{"id":1221,"date":"2026-10-23T04:21:16","date_gmt":"2026-10-23T04:21:16","guid":{"rendered":"https:\/\/getaiwork.com\/articles\/?p=1221"},"modified":"2026-10-24T04:17:23","modified_gmt":"2026-10-24T04:17:23","slug":"ai-researcher-jobs","status":"publish","type":"post","link":"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/","title":{"rendered":"AI Researcher Jobs: What 124 Lab Listings Ask For, Counted"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>AI researcher jobs on the labs&#8217; own boards on 23 October 2026 came to 124 research titled listings: 35 at OpenAI, 42 at Anthropic, 47 at Google DeepMind and none at xAI. 39 name a PhD, but only three state one with no alternative. The rest ask for a track record, years and code.<\/strong><\/p>\n\n\n\n<div class=\"msq-audio\" data-article=\"ai-researcher-jobs\" data-minutes=\"16\">\n<style>.msq-audio{margin:1.25em 0 1.75em;padding:16px 18px;background:#F1F4FA;border:1px solid #DDE3EE;border-radius:12px}\n.msq-audio-lead{margin:0 0 8px;font-size:.95em}.msq-audio audio{width:100%;margin:0 0 12px;display:block}\n.msq-toc-title{margin:0 0 6px;font-weight:700}.msq-toc ol{margin:0;padding:0;list-style:none}\n.msq-toc li{display:flex;gap:10px;align-items:center;padding:6px 6px;border-bottom:1px solid #DDE3EE;border-radius:8px}\n.msq-toc li:last-child{border-bottom:0}.msq-toc li.is-active{background:#fff}\n.msq-toc-play{flex:0 0 auto;width:30px;height:30px;border-radius:50%;border:0;background:#1E3A8A;color:#fff;font:inherit;font-size:.72em;line-height:30px;padding:0;cursor:pointer}\n.msq-toc-play.is-playing{background:#B45309}\n.msq-toc-time{flex:0 0 auto;min-width:3.1em;font-variant-numeric:tabular-nums;color:#55555F;font-size:.9em}\n.msq-toc a{color:#1B1F24;text-decoration:none;line-height:1.35}.msq-toc li.is-active a{font-weight:700;color:#1E3A8A}\n.msq-audio-note{margin:10px 0 0;font-size:.82em;color:#55555F}\nh2.wp-block-heading[id]{scroll-margin-top:24px}\n.msq-hplay{display:inline-block;vertical-align:middle;width:26px;height:26px;margin:0 0 0 10px;border-radius:50%;border:0;background:#1E3A8A;cursor:pointer;padding:0;position:relative;top:-2px}\n.msq-hplay::before{content:'\u25b6\ufe0e';display:block;color:#fff;font-size:11px;line-height:26px;text-align:center}\n.msq-hplay.is-playing{background:#B45309}.msq-hplay.is-playing::before{content:'\u2016';font-size:13px;font-weight:700}\n.msq-bar{position:fixed;left:0;right:0;bottom:0;z-index:9000;display:flex;align-items:center;gap:10px;padding:8px 12px;padding-bottom:calc(8px + env(safe-area-inset-bottom));background:#FFFFFF;border-top:1px solid #DDE3EE;box-shadow:0 -4px 16px rgba(0,0,0,.08);font-size:.92em;color:#1B1F24;transition:transform .2s ease}\n.msq-bar.is-hidden{transform:translateY(130%)}\n.msq-bar-play{display:inline-flex;align-items:center;gap:8px;border:0;border-radius:999px;background:#1E3A8A;color:#fff;font:inherit;font-weight:700;padding:8px 14px 8px 12px;cursor:pointer;white-space:nowrap}\n.msq-bar-play.is-playing{background:#B45309}\n.msq-bar-icon::before{content:'\u25b6\ufe0e';font-size:12px}.msq-bar-play.is-playing .msq-bar-icon::before{content:'\u2016';font-size:14px;font-weight:700}\n.msq-bar-chapter{flex:1 1 auto;min-width:0;overflow:hidden;text-overflow:ellipsis;white-space:nowrap}\n.msq-bar-time{flex:0 0 auto;font-variant-numeric:tabular-nums;color:#55555F;font-size:.9em}\n.msq-bar-rate,.msq-bar-close{flex:0 0 auto;border:1px solid #DDE3EE;background:#fff;color:#1B1F24;font:inherit;font-size:.85em;line-height:1;border-radius:8px;padding:7px 9px;cursor:pointer}\n.msq-bar-close::before{content:'\u2715'}\n.msq-bar.msq-bar-pill{left:auto;right:14px;bottom:14px;padding:6px;padding-bottom:6px;border:1px solid #DDE3EE;border-radius:999px;box-shadow:0 6px 20px rgba(0,0,0,.15)}\n.msq-bar.msq-bar-pill .msq-bar-chapter,.msq-bar.msq-bar-pill .msq-bar-time,.msq-bar.msq-bar-pill .msq-bar-rate{display:none}\nbody.msq-has-bar{padding-bottom:76px}\n@media (max-width:600px){.msq-bar-time{display:none}.msq-bar{gap:8px;padding-left:10px;padding-right:10px}}\n.msq-s.msq-now{background:#F1F4FA;box-shadow:0 2px 0 #1E3A8A;border-radius:2px}\ntr.msq-now td,tr.msq-now th{background:#F1F4FA;box-shadow:inset 0 -2px 0 #1E3A8A}\n@media (hover:hover){.msq-s,tr[data-i]{cursor:pointer}.msq-s:hover{background:#F1F4FA}}<\/style>\n<script type=\"application\/ld+json\">{\"@context\":\"https:\/\/schema.org\",\"@type\":\"AudioObject\",\"@id\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#audio\",\"name\":\"AI Researcher Jobs: What 124 Lab Listings Ask For, Counted, read aloud\",\"description\":\"Synthetic voice reading of the article, recorded 24 October 2026.\",\"contentUrl\":\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-audio-202610240947.mp3\",\"encodingFormat\":\"audio\/mpeg\",\"duration\":\"PT15M38S\",\"uploadDate\":\"2026-10-24\",\"inLanguage\":\"en-US\",\"isAccessibleForFree\":true,\"mainEntityOfPage\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/\",\"hasPart\":[{\"@type\":\"Clip\",\"name\":\"AI Researcher Jobs: What 124 Lab Listings Ask For, Counted\",\"startOffset\":0,\"endOffset\":54},{\"@type\":\"Clip\",\"name\":\"How many AI researcher jobs do the labs actually list?\",\"startOffset\":54,\"endOffset\":246,\"url\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#how-many-ai-researcher-jobs-do-the-labs-actually-list\"},{\"@type\":\"Clip\",\"name\":\"Do AI researcher jobs require a PhD?\",\"startOffset\":246,\"endOffset\":449,\"url\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#do-ai-researcher-jobs-require-a-phd\"},{\"@type\":\"Clip\",\"name\":\"What do the listings ask for instead of a PhD?\",\"startOffset\":449,\"endOffset\":569,\"url\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#what-do-the-listings-ask-for-instead-of-a-phd\"},{\"@type\":\"Clip\",\"name\":\"What do AI researcher jobs pay, as listed?\",\"startOffset\":569,\"endOffset\":752,\"url\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#what-do-ai-researcher-jobs-pay-as-listed\"},{\"@type\":\"Clip\",\"name\":\"What do people who got research roles without a PhD report?\",\"startOffset\":752,\"endOffset\":868,\"url\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#what-do-people-who-got-research-roles-without-a-phd-report\"},{\"@type\":\"Clip\",\"name\":\"How do you tell a real AI researcher job from the noise?\",\"startOffset\":868,\"endOffset\":938,\"url\":\"https:\/\/getaiwork.com\/articles\/ai-researcher-jobs\/#how-do-you-tell-a-real-ai-researcher-job-from-the-noise\"}],\"contentSize\":\"6.6 MB\"}<\/script>\n<p class=\"msq-audio-lead\"><strong>Listen to this article<\/strong>, about 16 minutes. A synthetic voice reads the article and announces each section title as it begins; tables are read row by row and the questions at the end are left out.<\/p>\n<audio class=\"msq-audio-player\" controls preload=\"metadata\" src=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-audio-202610240947.mp3\"><\/audio>\n<nav class=\"msq-toc\" aria-label=\"In this article\"><p class=\"msq-toc-title\">In this article<\/p>\n<ol>\n<li data-start=\"0.0\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"0.0\" aria-label=\"Play from: AI Researcher Jobs: What 124 Lab Listings Ask For, Counted\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">0:00<\/span><a href=\"#top\">AI Researcher Jobs: What 124 Lab Listings Ask For, Counted<\/a><\/li>\n<li data-start=\"53.58\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"53.58\" aria-label=\"Play from: How many AI researcher jobs do the labs actually list?\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">0:54<\/span><a href=\"#how-many-ai-researcher-jobs-do-the-labs-actually-list\">How many AI researcher jobs do the labs actually list?<\/a><\/li>\n<li data-start=\"245.88\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"245.88\" aria-label=\"Play from: Do AI researcher jobs require a PhD?\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">4:06<\/span><a href=\"#do-ai-researcher-jobs-require-a-phd\">Do AI researcher jobs require a PhD?<\/a><\/li>\n<li data-start=\"449.1\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"449.1\" aria-label=\"Play from: What do the listings ask for instead of a PhD?\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">7:29<\/span><a href=\"#what-do-the-listings-ask-for-instead-of-a-phd\">What do the listings ask for instead of a PhD?<\/a><\/li>\n<li data-start=\"569.16\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"569.16\" aria-label=\"Play from: What do AI researcher jobs pay, as listed?\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">9:29<\/span><a href=\"#what-do-ai-researcher-jobs-pay-as-listed\">What do AI researcher jobs pay, as listed?<\/a><\/li>\n<li data-start=\"752.38\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"752.38\" aria-label=\"Play from: What do people who got research roles without a PhD report?\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">12:32<\/span><a href=\"#what-do-people-who-got-research-roles-without-a-phd-report\">What do people who got research roles without a PhD report?<\/a><\/li>\n<li data-start=\"867.96\"><button type=\"button\" class=\"msq-toc-play\" data-start=\"867.96\" aria-label=\"Play from: How do you tell a real AI researcher job from the noise?\">&#x25B6;&#xFE0E;<\/button><span class=\"msq-toc-time\">14:28<\/span><a href=\"#how-do-you-tell-a-real-ai-researcher-job-from-the-noise\">How do you tell a real AI researcher job from the noise?<\/a><\/li>\n<\/ol><\/nav>\n<p class=\"msq-audio-note\">Recorded 24 October 2026 from the article as published that day. Press a chapter to jump there; the highlighted line is what is playing. A play button sits beside every section heading, and once you scroll past this player a bar at the bottom of the screen keeps the controls in reach. Tap or click any sentence to hear the reading start from there.<\/p>\n<script>var MSQ_SENT = [[0.0,\"h\",\"airesearcherjobswhat124l\",47],[4.8,\"p\",\"airesearcherjobsonthelab\",133],[22.48,\"p\",\"39nameaphdbutonlythreest\",47],[27.12,\"p\",\"therestaskforatrackrecor\",37],[31.64,\"p\",\"onthispageeveryresearcht\",223],[53.58,\"h\",\"howmanyairesearcherjobsd\",44],[58.06,\"p\",\"thepagesthatrankforthiss\",95],[63.58,\"p\",\"on23october2026theopenai\",135],[80.58,\"p\",\"apostingcountshereasares\",179],[94.42,\"p\",\"thatrulegives124research\",89],[103.62,\"p\",\"itsboardtitlesthatworkme\",145],[116.74,\"t\",\"openaiashby\",117],[130.7,\"t\",\"anthropicgreenhouse\",140],[147.14,\"t\",\"googledeepmindgooglecare\",143],[165.46,\"t\",\"xaigreenhouse\",120],[181.26,\"t\",\"cohereashby\",126],[200.14,\"p\",\"researchisathinsliceofev\",63],[205.82,\"p\",\"atopenai34ofthe35sitinsa\",48],[210.42,\"p\",\"anthropicspreadsits42acr\",114],[219.06,\"p\",\"deepminds47split28intheu\",87],[226.54,\"p\",\"thecountsofeveryanthropi\",95],[232.54,\"p\",\"metawhichismissingfromth\",141],[245.88,\"h\",\"doairesearcherjobsrequir\",29],[249.48,\"p\",\"readthelistingsandthehon\",65],[254.8,\"p\",\"acrossthe124researchlist\",57],[260.8,\"p\",\"36nameaphdandinthesamebr\",91],[267.44,\"p\",\"onlythreestateaphdundert\",86],[275.52,\"p\",\"thatis24oftheresearchlis\",44],[281.92,\"t\",\"openai\",91],[296.32,\"t\",\"anthropic\",94],[308.16,\"t\",\"googledeepmind\",100],[324.04,\"t\",\"xai\",86],[339.08,\"t\",\"fourlabstogether\",103],[352.6,\"t\",\"cohereforcomparison\",102],[367.84,\"p\",\"thewordingdiffersbylab\",22],[369.44,\"p\",\"openaissixphdsentencesal\",226],[390.6,\"p\",\"deepmindwritesoneclause2\",205],[410.72,\"p\",\"anthropicisthestrangeone\",24],[413.08,\"p\",\"all42researchlistingscar\",240],[432.56,\"p\",\"yettwolifesciencespostin\",133],[443.64,\"p\",\"bothlinesareontheposting\",77],[449.1,\"h\",\"whatdothelistingsaskfori\",36],[452.58,\"p\",\"the85listingswithnophdwo\",40],[455.98,\"p\",\"theyaskfortheevidenceado\",68],[461.1,\"p\",\"openaisplainresearchscie\",143],[471.46,\"p\",\"itsnewestresearchlisting\",137],[482.66,\"p\",\"trackrecordappearsinnine\",66],[488.86,\"p\",\"anthropicgoesfurtherthan\",31],[491.38,\"p\",\"fiveofitsresearchlisting\",242],[507.18,\"p\",\"thecertificateadvertprom\",135],[518.26,\"p\",\"deepminds20nonphdresearc\",259],[539.86,\"p\",\"oneresearchengineerpolar\",97],[548.46,\"p\",\"thepatternacrossallthree\",129],[557.62,\"p\",\"ifyouareweighingwhichoft\",165],[569.16,\"h\",\"whatdoairesearcherjobspa\",33],[572.52,\"p\",\"payhereisquotedonlyaspri\",55],[578.28,\"p\",\"listedbandschangeoffersd\",78],[584.0,\"p\",\"34ofopenais35researchlis\",168],[602.64,\"p\",\"anthropicprintsabandon40\",245],[631.0,\"p\",\"googlecareersprintsausba\",82],[637.88,\"t\",\"openai\",95],[652.8,\"t\",\"anthropic\",98],[665.84,\"t\",\"xaimemberoftechnicalstaf\",106],[679.28,\"t\",\"googledeepmind\",117],[689.24,\"t\",\"getaiworkboardstembackgr\",180],[717.36,\"p\",\"forscalethegetaiworkboar\",172],[739.84,\"p\",\"thelabresearchbandssitfa\",164],[752.38,\"h\",\"whatdopeoplewhogotresear\",48],[756.34,\"p\",\"twormachinelearningthrea\",67],[761.46,\"p\",\"inafebruary2026threadask\",130],[771.26,\"p\",\"onecommenterwithnophddes\",135],[781.9,\"p\",\"anotherwasrecruitedintoa\",161],[792.78,\"p\",\"athirdwhointerviewsresea\",98],[800.98,\"p\",\"apostdocsayseveryfriendn\",89],[808.14,\"p\",\"theapril2026threadisasta\",96],[817.9,\"p\",\"severalrepliessayapublic\",91],[824.46,\"p\",\"aformerresearchengineers\",81],[831.54,\"p\",\"onecommenterreportsgetti\",141],[841.58,\"p\",\"onedissentsthatthethread\",64],[846.18,\"p\",\"setagainstthelistingsthe\",139],[856.78,\"p\",\"thepostingsreadasopenthe\",73],[863.46,\"p\",\"thisiswhatpeoplereportno\",61],[867.96,\"h\",\"howdoyoutellarealairesea\",44],[871.2,\"p\",\"threetestswhichthe124lis\",51],[875.76,\"p\",\"thepostingsitsonthelabso\",117],[884.28,\"p\",\"itprintsapaybandinanamed\",32],[886.64,\"p\",\"anditsqualificationblock\",105],[895.44,\"p\",\"arolethatquotesaratenobo\",82],[903.4,\"p\",\"thefreepublicboardatgeta\",164],[915.36,\"p\",\"labcountsmoveweeklyanthr\",209]];\n(function(){\n  function init(){\n  var root = document.querySelector('.msq-audio'); if (!root) { return; }\n  var audio = root.querySelector('audio'), items = Array.prototype.slice.call(root.querySelectorAll('.msq-toc li'));\n  var starts = items.map(function(li){ return parseFloat(li.getAttribute('data-start')) || 0; });\n  var names = items.map(function(li){ return li.querySelector('a').textContent; });\n  var anchors = items.map(function(li){ var h = li.querySelector('a').getAttribute('href') || ''; return (h.charAt(0) === '#' && h !== '#top') ? document.getElementById(h.slice(1)) : null; });\n  var article = root.getAttribute('data-article') || location.pathname, minutes = root.getAttribute('data-minutes') || '';\n  var marks = {}, started = false, pausedAt = -1, closed = false, ended = false;\n  var RATES = [1, 1.25, 1.5], rate = 1;\n  try { var saved = parseFloat(localStorage.getItem('msq-audio-rate')); if (RATES.indexOf(saved) >= 0) { rate = saved; } } catch (e) {}\n  audio.playbackRate = rate;\n  function track(name, extra){ try { if (typeof gtag === 'function') { var p = { article: article }; for (var k in extra) { p[k] = extra[k]; } gtag('event', name, p); } } catch (e) {} }\n  function current(){ var t = audio.currentTime, idx = 0; for (var i = 0; starts.length > i; i += 1) { if (t + 0.25 >= starts[i]) { idx = i; } } return idx; }\n  function onScreen(){\n    var line = window.innerHeight \/ 3, idx = 0;\n    for (var i = 0; anchors.length > i; i += 1) { var el = anchors[i]; if (el && line >= el.getBoundingClientRect().top) { idx = i; } }\n    return idx;\n  }\n  function fmt(s){ s = Math.max(0, Math.round(s || 0)); var m = Math.floor(s \/ 60), r = s % 60; return m + ':' + (10 > r ? '0' : '') + r; }\n  function play(from, source){\n    if (from !== null) { audio.currentTime = from; }\n    var p = audio.play(); if (p && p.catch) { p.catch(function(){}); }\n    track(from === null ? 'audio_resume' : 'audio_chapter', { chapter: names[current()], start: Math.round(audio.currentTime), source: source });\n  }\n  var hbtns = anchors.map(function(el, i){\n    if (!el) { return null; }\n    var b = document.createElement('button'); b.type = 'button'; b.className = 'msq-hplay'; b.setAttribute('aria-label', 'Play this section: ' + names[i]);\n    b.addEventListener('click', function(){\n      if (!audio.paused && current() === i) { audio.pause(); track('audio_pause', { source: 'heading' }); return; }\n      play(starts[i], 'heading');\n    });\n    el.appendChild(b); return b;\n  });\n  var bar = document.createElement('div'); bar.className = 'msq-bar msq-bar-pill is-hidden'; bar.setAttribute('role', 'region'); bar.setAttribute('aria-label', 'Audio player');\n  var playBtn = document.createElement('button'); playBtn.type = 'button'; playBtn.className = 'msq-bar-play';\n  var icon = document.createElement('span'); icon.className = 'msq-bar-icon'; icon.setAttribute('aria-hidden', 'true');\n  var label = document.createElement('span'); label.className = 'msq-bar-label';\n  playBtn.appendChild(icon); playBtn.appendChild(label);\n  var chap = document.createElement('span'); chap.className = 'msq-bar-chapter';\n  var clock = document.createElement('span'); clock.className = 'msq-bar-time';\n  var rateBtn = document.createElement('button'); rateBtn.type = 'button'; rateBtn.className = 'msq-bar-rate'; rateBtn.setAttribute('aria-label', 'Playback speed');\n  var closeBtn = document.createElement('button'); closeBtn.type = 'button'; closeBtn.className = 'msq-bar-close'; closeBtn.setAttribute('aria-label', 'Close the audio bar');\n  bar.appendChild(playBtn); bar.appendChild(chap); bar.appendChild(clock); bar.appendChild(rateBtn); bar.appendChild(closeBtn);\n  document.body.appendChild(bar);\n  function showRate(){ rateBtn.textContent = (rate === 1 ? '1' : String(rate)) + 'x'; }\n  showRate();\n  rateBtn.addEventListener('click', function(){\n    rate = RATES[(RATES.indexOf(rate) + 1) % RATES.length]; audio.playbackRate = rate; showRate();\n    try { localStorage.setItem('msq-audio-rate', String(rate)); } catch (e) {}\n    track('audio_rate', { rate: rate });\n  });\n  closeBtn.addEventListener('click', function(){\n    closed = true; if (!audio.paused) { audio.pause(); } paint(); track('audio_bar_close', {});\n  });\n  function mode(){\n    if (!audio.paused) { return 'pause'; }\n    if (ended) { return 'again'; }\n    if (!started) { return 'listen'; }\n    return (pausedAt === onScreen()) ? 'resume' : 'section';\n  }\n  playBtn.addEventListener('click', function(){\n    var m = mode();\n    if (m === 'pause') { audio.pause(); track('audio_pause', { source: 'bar' }); return; }\n    if (m === 'again') { ended = false; play(0, 'bar'); return; }\n    if (m === 'resume') { play(null, 'bar'); return; }\n    play(starts[onScreen()], 'bar');\n  });\n  function paint(){\n    var idx = current(), playing = !audio.paused, m = mode();\n    items.forEach(function(li, i){\n      var on = (i === idx) && playing, b = li.querySelector('.msq-toc-play'), a = li.querySelector('a');\n      li.classList.toggle('is-active', on); b.classList.toggle('is-playing', on);\n      b.textContent = on ? '\u2016' : '\u25b6\ufe0e';\n      b.setAttribute('aria-label', (on ? 'Pause: ' : 'Play from: ') + a.textContent);\n    });\n    hbtns.forEach(function(b, i){ if (b) { var on = (i === idx) && playing; b.classList.toggle('is-playing', on); b.setAttribute('aria-label', (on ? 'Pause this section: ' : 'Play this section: ') + names[i]); } });\n    playBtn.classList.toggle('is-playing', playing);\n    label.textContent = m === 'pause' ? 'Pause' : m === 'again' ? 'Play again' : m === 'resume' ? 'Resume' : m === 'section' ? 'Play this section' : 'Listen, ' + minutes + ' min';\n    playBtn.setAttribute('aria-label', label.textContent);\n    chap.textContent = started ? names[(playing || m === 'resume') ? idx : onScreen()] : '';\n    clock.textContent = started ? fmt(audio.currentTime) + ' \/ ' + fmt(audio.duration) : '';\n    bar.classList.toggle('msq-bar-pill', !started);\n    var show = !closed && (started || 0 > root.getBoundingClientRect().bottom);\n    bar.classList.toggle('is-hidden', !show);\n    document.body.classList.toggle('msq-has-bar', show);\n  }\n  root.addEventListener('click', function(e){\n    var b = e.target.closest ? e.target.closest('.msq-toc-play') : null; if (!b) { return; }\n    var li = b.parentNode, s = parseFloat(b.getAttribute('data-start')) || 0;\n    if (!audio.paused && current() === items.indexOf(li)) { audio.pause(); track('audio_pause', { source: 'index' }); return; }\n    play(s, 'index');\n  });\n  audio.addEventListener('play', function(){ ended = false; closed = false; if (!started) { started = true; track('audio_play', { minutes: Math.round((audio.duration || 0) \/ 60) }); } paint(); });\n  audio.addEventListener('pause', function(){ pausedAt = current(); paint(); });\n  audio.addEventListener('timeupdate', function(){\n    paint();\n    if (audio.duration > 0) {\n      var pct = Math.floor(audio.currentTime \/ audio.duration * 100);\n      [25, 50, 75].forEach(function(m){ if (pct >= m && !marks[m]) { marks[m] = true; track('audio_progress', { percent: m }); } });\n    }\n  });\n  audio.addEventListener('ended', function(){ ended = true; track('audio_complete', {}); paint(); });\n  var ticking = false;\n  function onScroll(){ if (ticking) { return; } ticking = true; setTimeout(function(){ ticking = false; paint(); }, 80); }\n  window.addEventListener('scroll', onScroll, { passive: true }); window.addEventListener('resize', onScroll);\n  paint();\n  }\n  if (document.readyState === 'loading') { document.addEventListener('DOMContentLoaded', init); } else { init(); }\n})();\n\n(function(){\n  function start(){\n    var root = document.querySelector('.msq-audio'), audio = root ? root.querySelector('audio') : null;\n    var map = (typeof MSQ_SENT !== 'undefined') ? MSQ_SENT : [];\n    if (!root || !audio || !map.length) { return; }\n    var scope = root.closest('main') || root.closest('article') || root.closest('.entry-content') || document.body;\n    var SKIP = '.msq-audio,.msq,.msq-tool,.msq-end,.msq-bar,script,style,noscript,figcaption,#rank-math-faq,.rank-math-block,nav,header,footer';\n    function norm(s){ return s.toLowerCase().replace(\/[^a-z0-9]+\/g, ''); }\n    var nodes = [], stream = '', walker = document.createTreeWalker(scope, NodeFilter.SHOW_TEXT, null), n;\n    while ((n = walker.nextNode())) {\n      var el = n.parentElement; if (!el || el.closest(SKIP)) { continue; }\n      var raw = n.nodeValue, keep = [], norms = '';\n      for (var i = 0; raw.length > i; i += 1) { var c = raw.charAt(i).toLowerCase(); if (\/[a-z0-9]\/.test(c)) { keep.push(i); norms += c; } }\n      if (!norms) { continue; }\n      nodes.push({ node: n, a: stream.length, b: stream.length + norms.length, keep: keep });\n      stream += norms;\n    }\n    var segs = [], cursor = 0, lastRow = null;\n    map.forEach(function(m, idx){\n      var kind = m[1], f = m[2], len = m[3];\n      if (kind === 't') {\n        var rows = scope.querySelectorAll('tr'), after = !lastRow;\n        for (var r = 0; rows.length > r; r += 1) {\n          var tr = rows[r];\n          if (!after) { if (tr === lastRow) { after = true; } continue; }\n          var cell = tr.querySelector('td,th');\n          if (cell && norm(cell.textContent).indexOf(f) === 0) { tr.setAttribute('data-i', idx); lastRow = tr; break; }\n        }\n        return;\n      }\n      var p = stream.indexOf(f, cursor); if (0 > p) { return; }\n      segs.push({ i: idx, a: p, b: Math.min(p + len, stream.length) }); cursor = p + len;\n    });\n    var si = 0;\n    nodes.forEach(function(nd){\n      var frag = document.createDocumentFragment(), raw = nd.node.nodeValue, rawPos = 0, any = false;\n      for (var k = si; segs.length > k; k += 1) {\n        var sg = segs[k]; if (sg.a >= nd.b) { break; } if (nd.a >= sg.b) { continue; }\n        var s = Math.max(sg.a, nd.a), e = Math.min(sg.b, nd.b);\n        var rs = nd.keep[s - nd.a], re2 = nd.keep[e - nd.a - 1] + 1;\n        if (rs > rawPos) { frag.appendChild(document.createTextNode(raw.slice(rawPos, rs))); }\n        var span = document.createElement('span'); span.className = 'msq-s'; span.setAttribute('data-i', sg.i); span.textContent = raw.slice(rs, re2);\n        frag.appendChild(span); rawPos = re2; any = true;\n        if (sg.b > nd.b) { break; }\n      }\n      while (segs.length > si && nd.b >= segs[si].b) { si += 1; }\n      if (!any) { return; }\n      if (raw.length > rawPos) { frag.appendChild(document.createTextNode(raw.slice(rawPos))); }\n      nd.node.parentNode.replaceChild(frag, nd.node);\n    });\n    var starts = map.map(function(m){ return m[0]; }), curIdx = -1, lastUser = 0, timer = null, autoUntil = 0, dragging = false;\n    function idxAt(t){ var lo = 0, hi = starts.length; while (hi > lo + 1) { var mid = (lo + hi) >> 1; if (t + 0.15 >= starts[mid]) { lo = mid; } else { hi = mid; } } return (t + 0.15 >= starts[0]) ? lo : -1; }\n    function setNow(i){\n      if (i === curIdx) { return; }\n      Array.prototype.forEach.call(scope.querySelectorAll('.msq-now'), function(el){ el.classList.remove('msq-now'); });\n      curIdx = i; if (0 > i) { return; }\n      var els = scope.querySelectorAll('[data-i=\"' + i + '\"]'); if (!els.length) { return; }\n      Array.prototype.forEach.call(els, function(el){ el.classList.add('msq-now'); });\n      var r = els[0].getBoundingClientRect(), now = Date.now();\n      if (!dragging && now > lastUser + 2500 && (80 > r.top || r.bottom > window.innerHeight - 140)) {\n        var reduce = window.matchMedia && window.matchMedia('(prefers-reduced-motion: reduce)').matches;\n        autoUntil = now + (reduce ? 400 : 1600);\n        els[0].scrollIntoView({ block: 'center', behavior: reduce ? 'auto' : 'smooth' });\n      }\n    }\n    function tick(){ if (!audio.paused) { setNow(idxAt(audio.currentTime)); } }\n    function userScrolled(){ lastUser = Date.now(); }\n    ['wheel', 'touchmove', 'keydown'].forEach(function(ev){ window.addEventListener(ev, userScrolled, { passive: true }); });\n    window.addEventListener('scroll', function(){ if (Date.now() > autoUntil) { userScrolled(); } }, { passive: true });\n    window.addEventListener('mousedown', function(e){ if (e.clientX >= document.documentElement.clientWidth) { dragging = true; userScrolled(); } }, { passive: true });\n    window.addEventListener('mouseup', function(){ if (dragging) { dragging = false; userScrolled(); } }, { passive: true });\n    audio.addEventListener('play', function(){ if (timer) { clearInterval(timer); } timer = setInterval(tick, 150); tick(); });\n    audio.addEventListener('pause', function(){ if (timer) { clearInterval(timer); timer = null; } });\n    scope.addEventListener('click', function(e){\n      if (e.target.closest('a, button, input, select, textarea, label, summary, .msq, .msq-tool, .msq-end, .msq-audio, .msq-bar')) { return; }\n      var sel = window.getSelection ? String(window.getSelection()) : '';\n      if (sel && sel.trim()) { return; }\n      var el = e.target.closest('[data-i]'); if (!el) { return; }\n      var i = parseInt(el.getAttribute('data-i'), 10); if (isNaN(i)) { return; }\n      if (!audio.paused && i === curIdx) { return; }\n      lastUser = Date.now();\n      audio.currentTime = starts[i];\n      var p = audio.play(); if (p && p.catch) { p.catch(function(){}); }\n      try { if (typeof gtag === 'function') { gtag('event', 'audio_chapter', { article: location.pathname, source: 'sentence', start: Math.round(starts[i]) }); } } catch (err) {}\n    });\n    audio.addEventListener('seeked', tick);\n    audio.addEventListener('ended', function(){ setNow(-1); });\n  }\n  if (document.readyState === 'loading') { document.addEventListener('DOMContentLoaded', start); } else { start(); }\n})();<\/script>\n<\/div>\n\n\n\n<p class=\"msq-hook wp-block-paragraph\">On this page: every research titled listing on the OpenAI, Anthropic, Google DeepMind and xAI boards counted on 23 October 2026, the PhD line read on each of the 124, pay only as printed on the posting, and what two 2026 forum threads report about getting in without a doctorate.<\/p>\n\n\n\n<div class=\"msq\" id=\"ms-quiz\" data-slug=\"ai-researcher-jobs\" data-category=\"ai-jobs\" data-list=\"1\">\n<style>.msq,.msq-hook,.msq-box,.msq-end,.msq-tool{--msq-accent:#1E3A8A;--msq-accent-text:#FFFFFF;--msq-line:#DDE3EE;--msq-warn:#B45309;--msq-muted:#55555F;--msq-tint:#F1F4FA;--msq-base:#FFFFFF}\n.msq{margin:1.25em 0 1.75em}\n.msq-hook{border-left:3px solid var(--msq-accent);padding:6px 14px;color:var(--msq-muted);font-size:.97em}\n.msq-box{background:var(--msq-tint);border:1px solid var(--msq-line);border-radius:12px;padding:0}\n.msq-box>summary{cursor:pointer;list-style:none;padding:14px 18px;font-weight:700;color:var(--msq-accent)}\n.msq-box>summary::-webkit-details-marker{display:none}\n.msq-box>summary::before{content:\"+\";display:inline-block;width:1.1em;font-weight:800}\n.msq-box[open]>summary::before{content:\"\\00d7\"}\n.msq-body{padding:2px 18px 18px}\n.msq-intro{margin:0 0 12px;font-size:.95em;color:var(--msq-muted)}\n.msq-list{margin:0;padding:0;list-style:none}\n.msq-q{margin:0 0 18px;padding:0}\n.msq-question{font-weight:700;margin:0 0 8px}\n.msq-options{display:flex;flex-direction:column;gap:8px}\n.msq-opt{text-align:left;background:var(--msq-base);border:1px solid var(--msq-line);border-radius:8px;padding:10px 12px;font:inherit;font-size:.95em;color:inherit;cursor:pointer;line-height:1.35}\n.msq-opt:hover{border-color:var(--msq-accent)}\n.msq-opt[disabled]{cursor:default}\n.msq-opt.is-right{border-color:var(--msq-accent);background:var(--msq-line);font-weight:700}\n.msq-opt.is-wrong{border-color:var(--msq-warn);color:var(--msq-warn)}\n.msq-explain{margin:8px 0 0;font-size:.92em;padding:10px 12px;border-left:3px solid var(--msq-accent);background:var(--msq-base);border-radius:0 8px 8px 0}\n.msq-explain[hidden]{display:none}\n.msq-verdict{font-weight:700}\n.msq-result{margin-top:14px;padding:14px 16px;background:var(--msq-base);border:1px solid var(--msq-line);border-radius:10px}\n.msq-result[hidden]{display:none}\n.msq-result h3{margin:0 0 6px;font-size:1.1em}\n.msq-result p{margin:0 0 8px}\n.msq-jump{margin:0 0 8px;padding-left:1.2em}\n.msq-form{margin-top:10px;display:flex;flex-wrap:wrap;gap:8px;align-items:center}\n.msq-form input[type=email]{flex:1 1 220px;padding:10px 12px;border:1px solid var(--msq-muted);border-radius:8px;font:inherit;background:#fff}\n.msq-btn{background:var(--msq-accent);color:var(--msq-accent-text);border:0;border-radius:8px;padding:11px 18px;font:inherit;font-weight:700;cursor:pointer}\n.msq-btn:hover{filter:brightness(.92)}\n.msq-btn[disabled]{opacity:.7;cursor:default}\n.msq-no{font-size:.9em;color:var(--msq-muted);margin-left:4px}\n.msq-privacy{margin:6px 0 0;font-size:.82em;color:var(--msq-muted);flex-basis:100%}\n.msq-hp{position:absolute;left:-5000px;width:1px;height:1px;overflow:hidden}\n.msq-msg{margin:8px 0 0;font-size:.92em}\n.msq-end{margin:1.5em 0;padding:18px;border:1px solid var(--msq-accent);border-radius:12px;text-align:center;background:var(--msq-tint)}\n.msq-end p{margin:0 0 12px}\n.msq-end .msq-body{text-align:left;padding:16px 0 0}\n.msq-end .msq-start-test[hidden]{display:none}\n.msq-tool{margin:1.5em 0;padding:18px 20px;background:var(--msq-tint);border:1px solid var(--msq-line);border-radius:12px}\n.msq-tool h3{margin:0 0 6px;font-size:1.15em}\n.msq-tool p{margin:0 0 10px}\n.msq-tool .msq-note{font-size:.85em;color:var(--msq-muted);margin:10px 0 0}\n.msq-check label{display:flex;gap:10px;align-items:flex-start;padding:9px 0;border-bottom:1px solid var(--msq-line);cursor:pointer;line-height:1.4}\n.msq-check label:last-of-type{border-bottom:0}\n.msq-check input{margin-top:5px;flex:0 0 auto;width:18px;height:18px;accent-color:var(--msq-accent)}\n.msq-check-count{font-weight:700;margin:10px 0 0}\n.msq-verdict-box{margin:12px 0 0;padding:10px 14px;border-radius:8px;background:var(--msq-base);border-left:3px solid var(--msq-warn)}\n.msq-verdict-box[hidden]{display:none}\n.msq-calc{display:grid;grid-template-columns:1fr 1fr;gap:10px 16px}\n.msq-calc label{display:flex;flex-direction:column;gap:4px;font-size:.92em}\n.msq-calc input,.msq-calc select{padding:8px 10px;border:1px solid var(--msq-muted);border-radius:8px;font:inherit;background:#fff}\n.msq-calc-out{grid-column:1 \/ -1;margin-top:6px;padding:12px 14px;background:var(--msq-base);border-radius:8px;border:1px solid var(--msq-line)}\n.msq-calc-out strong{font-size:1.15em}\n.msq-choose label{display:block;padding:9px 12px;margin:0 0 8px;background:var(--msq-base);border:1px solid var(--msq-line);border-radius:8px;cursor:pointer}\n.msq-choose input{margin-right:8px;accent-color:var(--msq-accent)}\n.msq-choice-result{margin-top:10px;padding:12px 14px;background:var(--msq-base);border-left:3px solid var(--msq-accent);border-radius:0 8px 8px 0}\n.msq-choice-result[hidden]{display:none}\n@media (max-width:600px){.msq-calc{grid-template-columns:1fr}}\nh2.wp-block-heading[id]{scroll-margin-top:24px}<\/style>\n<noscript><style>.msq-explain{display:block !important}<\/style><\/noscript>\n<details class=\"msq-box\">\n<summary>Think you already know this? Five questions, one minute.<\/summary>\n<div class=\"msq-body\">\n<p class=\"msq-intro\">Every answer is on this page. Answer from what you know now; the page fills in the rest.<\/p>\n<ol class=\"msq-list\">\n<li class=\"msq-q\" data-q=\"1\" data-answer=\"a\" data-title=\"How many AI researcher jobs do the labs actually list?\" data-section=\"#how-many-ai-researcher-jobs-do-the-labs-actually-list\">\n<p class=\"msq-question\">1. How many research titled listings did the OpenAI, Anthropic, Google DeepMind and xAI boards carry between them on 23 October 2026?<\/p>\n<div class=\"msq-options\" role=\"group\" aria-label=\"Question 1 options\"><button type=\"button\" class=\"msq-opt\" data-opt=\"a\">124<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"b\">802<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"c\">35<\/button><\/div>\n<p class=\"msq-explain\" hidden><span class=\"msq-verdict\"><\/span>35 at OpenAI, 42 at Anthropic, 47 at Google DeepMind and none at xAI, counted by title on each lab&#8217;s own board, starting with the <a href=\"https:\/\/jobs.ashbyhq.com\/openai\" target=\"_blank\" rel=\"noopener\">OpenAI board<\/a> and its 802 postings. <a class=\"msq-section\" href=\"#how-many-ai-researcher-jobs-do-the-labs-actually-list\">Read the section<\/a><\/p>\n<\/li>\n<li class=\"msq-q\" data-q=\"2\" data-answer=\"b\" data-title=\"Do AI researcher jobs require a PhD?\" data-section=\"#do-ai-researcher-jobs-require-a-phd\">\n<p class=\"msq-question\">2. How many of the 124 research listings stated a PhD with no alternative on 23 October 2026?<\/p>\n<div class=\"msq-options\" role=\"group\" aria-label=\"Question 2 options\"><button type=\"button\" class=\"msq-opt\" data-opt=\"a\">39<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"b\">3<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"c\">85<\/button><\/div>\n<p class=\"msq-explain\" hidden><span class=\"msq-verdict\"><\/span>Only three: one DeepMind role and two Anthropic life sciences roles such as <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/5285250008\" target=\"_blank\" rel=\"noopener\">Research Scientist, Life Sciences (Experimental Biology)<\/a>. 39 named a PhD at all and 85 never mentioned one. <a class=\"msq-section\" href=\"#do-ai-researcher-jobs-require-a-phd\">Read the section<\/a><\/p>\n<\/li>\n<li class=\"msq-q\" data-q=\"3\" data-answer=\"a\" data-title=\"What do the listings ask for instead of a PhD?\" data-section=\"#what-do-the-listings-ask-for-instead-of-a-phd\">\n<p class=\"msq-question\">3. What do five Anthropic research listings say strong candidates need not have?<\/p>\n<div class=\"msq-options\" role=\"group\" aria-label=\"Question 3 options\"><button type=\"button\" class=\"msq-opt\" data-opt=\"a\">Formal education credentials or a publication history<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"b\">Experience with Python<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"c\">A United States work permit<\/button><\/div>\n<p class=\"msq-explain\" hidden><span class=\"msq-verdict\"><\/span>The list headed strong candidates need not have on postings such as <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/4613568008\" target=\"_blank\" rel=\"noopener\">Research Engineer, Machine Learning (Reinforcement Learning)<\/a> opens with formal certifications or education credentials and academic research experience or publication history. <a class=\"msq-section\" href=\"#what-do-the-listings-ask-for-instead-of-a-phd\">Read the section<\/a><\/p>\n<\/li>\n<li class=\"msq-q\" data-q=\"4\" data-answer=\"b\" data-title=\"What do AI researcher jobs pay, as listed?\" data-section=\"#what-do-ai-researcher-jobs-pay-as-listed\">\n<p class=\"msq-question\">4. What was the highest ceiling printed on an OpenAI or Anthropic research listing on 23 October 2026?<\/p>\n<div class=\"msq-options\" role=\"group\" aria-label=\"Question 4 options\"><button type=\"button\" class=\"msq-opt\" data-opt=\"a\">$445,000<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"b\">$850,000<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"c\">$1,100,000<\/button><\/div>\n<p class=\"msq-explain\" hidden><span class=\"msq-verdict\"><\/span>Both boards topped out at $850,000 as listed that day, for example the <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/5459098008\" target=\"_blank\" rel=\"noopener\">Research Engineer, Chip Design RL<\/a> posting at $500,000 to $850,000. Listed bands change and nothing is guaranteed. <a class=\"msq-section\" href=\"#what-do-ai-researcher-jobs-pay-as-listed\">Read the section<\/a><\/p>\n<\/li>\n<li class=\"msq-q\" data-q=\"5\" data-answer=\"a\" data-title=\"What do people who got research roles without a PhD report?\" data-section=\"#what-do-people-who-got-research-roles-without-a-phd-report\">\n<p class=\"msq-question\">5. In the February 2026 r\/MachineLearning thread, what did the commenter hired by Amazon as a research scientist with a bachelor&#8217;s degree credit for the offer?<\/p>\n<div class=\"msq-options\" role=\"group\" aria-label=\"Question 5 options\"><button type=\"button\" class=\"msq-opt\" data-opt=\"a\">Client work at a small startup, and timing<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"b\">An AI certificate<\/button><button type=\"button\" class=\"msq-opt\" data-opt=\"c\">A cold online application<\/button><\/div>\n<p class=\"msq-explain\" hidden><span class=\"msq-verdict\"><\/span>That account in the <a href=\"https:\/\/www.reddit.com\/r\/MachineLearning\/comments\/1r0xtzd\/d_for_those_of_you_who_secured_research_scientist\/\" target=\"_blank\" rel=\"noopener\">February 2026 thread<\/a> credits industry work for a healthcare client and an emerging business need, and adds that an online application today would not get a look. It is what one person reports, not a measurement. <a class=\"msq-section\" href=\"#what-do-people-who-got-research-roles-without-a-phd-report\">Read the section<\/a><\/p>\n<\/li>\n<\/ol>\n<div class=\"msq-result\" hidden><\/div>\n<\/div>\n<\/details>\n<script>(function(){\n  if (window.__msqInit) { return; }\n  window.__msqInit = true;\n  var root = document.getElementById('ms-quiz');\n  if (!root) { return; }\n  var meta = { slug: root.getAttribute('data-slug'), category: root.getAttribute('data-category'),\n               list: root.getAttribute('data-list') || '1' };\n  function qs(el, s){ return el.querySelector(s); }\n  function qsa(el, s){ return Array.prototype.slice.call(el.querySelectorAll(s)); }\n  function track(name, params){\n    try { if (typeof gtag === 'function') { var p = params || {}; p.article = meta.slug; gtag('event', name, p); } } catch (e) {}\n  }\n  function jumpList(missed){\n    if (!missed.length) { return ''; }\n    return '<ul class=\"msq-jump\">' + missed.map(function(q){\n      return '<li><a href=\"' + q.getAttribute('data-section') + '\">' + q.getAttribute('data-title') + '<\/a><\/li>';\n    }).join('') + '<\/ul>';\n  }\n  function emailForm(){\n    return '<div class=\"msq-sub\"><p><strong>Get the next one.<\/strong> We publish new articles on ' + meta.category\n      + ' most weeks and recheck this page on Sundays. Email me when a new article goes up or this page changes.<\/p>'\n      + '<form class=\"msq-form\" method=\"post\" action=\"\/mail\/form.php\" novalidate>'\n      + '<input type=\"hidden\" name=\"list\" value=\"' + meta.list + '\">'\n      + '<input type=\"hidden\" name=\"name\" value=\"\">'\n      + '<label class=\"msq-hp\" aria-hidden=\"true\">Leave this empty<input type=\"text\" name=\"website\" tabindex=\"-1\" autocomplete=\"off\"><\/label>'\n      + '<input type=\"email\" name=\"email\" placeholder=\"you@example.com\" required aria-label=\"Email address\" autocomplete=\"email\">'\n      + '<button type=\"submit\" class=\"msq-btn\">Email me the updates<\/button>'\n      + '<a href=\"#\" class=\"msq-no\">No thanks<\/a>'\n      + '<p class=\"msq-privacy\">We store your email and this page\\'s topic, nothing else. Unsubscribe any time.<\/p>'\n      + '<\/form><p class=\"msq-msg\" aria-live=\"polite\"><\/p><\/div>';\n  }\n  \/* The autoresponder always answers 200 with its own page. A \".msg.err\" box means the\n     address was rejected, \"You're subscribed\" means accepted, and a redirect only happens\n     after a successful subscribe. Same reading as the landing pages on the other sites. *\/\n  function sendForm(form){\n    return fetch(form.getAttribute('action'), { method: 'POST', body: new FormData(form), credentials: 'same-origin', redirect: 'manual' })\n      .then(function(res){\n        if (res.type === 'opaqueredirect') { return { status: 'ok' }; }\n        if (res.status === 404) { return { status: 'notlive' }; }\n        if (!res.ok) { throw new Error('HTTP ' + res.status); }\n        return res.text().then(function(t){\n          var doc = new DOMParser().parseFromString(t, 'text\/html');\n          var err = doc.querySelector('.msg.err');\n          if (err) {\n            var m = err.textContent.trim();\n            if (\/something went wrong\/i.test(m)) { throw new Error(m); }\n            return { status: 'rejected', message: m };\n          }\n          if (\/you.re subscribed\/i.test(doc.body.textContent)) { return { status: 'ok' }; }\n          return { status: 'unknown' };\n        });\n      });\n  }\n  function wireForm(sub){\n    var form = qs(sub, '.msq-form'), msg = qs(sub, '.msq-msg');\n    qs(form, '.msq-no').addEventListener('click', function(e){\n      e.preventDefault();\n      sub.innerHTML = '<p class=\"msq-msg\">No problem. The Sunday refresh keeps this page current either way.<\/p>';\n      track('quiz_email', { result: 'declined' });\n    });\n    form.addEventListener('submit', function(e){\n      e.preventDefault();\n      var input = qs(form, 'input[name=email]'), btn = qs(form, '.msq-btn');\n      var email = input.value.trim(); input.value = email;\n      if (!\/^[^\\s@]+@[^\\s@]+\\.[^\\s@]+$\/.test(email)) { msg.textContent = 'That email does not look right.'; return; }\n      btn.disabled = true; btn.textContent = 'Sending';\n      function reset(){ btn.disabled = false; btn.textContent = 'Email me the updates'; }\n      sendForm(form).then(function(r){\n        if (r.status === 'ok' || r.status === 'unknown') {\n          sub.innerHTML = '<p class=\"msq-msg\"><strong>Done.<\/strong> We will email you when the next article on ' + meta.category\n            + ' goes up and when this page changes. If a confirmation email arrives, click it.<\/p>';\n          track('quiz_email', { result: r.status });\n          if (r.status === 'ok') { track('generate_lead', { method: 'quiz' }); }\n        } else if (r.status === 'notlive') {\n          msg.textContent = 'Sign up is not switched on yet for this site. Nothing was stored.'; reset();\n          track('quiz_email', { result: 'notlive' });\n        } else {\n          msg.textContent = r.message || 'Please check the email address and try again.'; reset();\n          track('quiz_email', { result: 'rejected' });\n        }\n      }).catch(function(){\n        msg.textContent = 'Could not reach the sign up service just now. Nothing was stored.'; reset();\n        track('quiz_email', { result: 'error' });\n      });\n    });\n  }\n  function finish(container, mode, score, total, missed){\n    var result = qs(container, '.msq-result');\n    track('quiz_complete', { mode: mode, score: score, total: total });\n    var h;\n    if (mode === 'pre') {\n      h = '<h3>You knew ' + score + ' of ' + total + ' before reading.<\/h3>';\n      h += (score === total) ? '<p>All of them. Read on anyway; the sections below carry the detail behind each answer.<\/p>'\n                             : '<p>The rest is below. Jump straight to what you missed:<\/p>' + jumpList(missed);\n      try { sessionStorage.setItem('msq-pre-' + meta.slug, String(score)); } catch (e) {}\n    } else {\n      var pre = null;\n      try { pre = sessionStorage.getItem('msq-pre-' + meta.slug); } catch (e) {}\n      h = '<h3>You got ' + score + ' of ' + total + '.<\/h3>';\n      if (pre !== null) { h += '<p>Before reading you had ' + pre + ' of ' + total + '.<\/p>'; }\n      if (missed.length) { h += '<p>Worth a second look:<\/p>' + jumpList(missed); }\n      h += emailForm();\n    }\n    result.innerHTML = h; result.hidden = false;\n    if (mode !== 'pre') { wireForm(qs(result, '.msq-sub')); }\n    result.scrollIntoView({ behavior: 'smooth', block: 'nearest' });\n  }\n  function setup(container, mode){\n    var qEls = qsa(container, '.msq-q'), answered = 0, score = 0, missed = [];\n    qEls.forEach(function(q){\n      var answer = q.getAttribute('data-answer'), opts = qsa(q, '.msq-opt'), explain = qs(q, '.msq-explain');\n      q.removeAttribute('data-done'); explain.hidden = true; qs(explain, '.msq-verdict').textContent = '';\n      opts.forEach(function(b){ b.disabled = false; b.classList.remove('is-right', 'is-wrong'); });\n      opts.forEach(function(b){\n        b.addEventListener('click', function(){\n          if (q.getAttribute('data-done')) { return; }\n          q.setAttribute('data-done', '1');\n          var ok = b.getAttribute('data-opt') === answer;\n          opts.forEach(function(o){ o.disabled = true; if (o.getAttribute('data-opt') === answer) { o.classList.add('is-right'); } });\n          if (ok) { score += 1; } else { b.classList.add('is-wrong'); missed.push(q); }\n          qs(explain, '.msq-verdict').textContent = ok ? 'Right. ' : 'Not quite. ';\n          explain.hidden = false;\n          answered += 1;\n          track('quiz_answer', { mode: mode, question: q.getAttribute('data-q'), correct: ok ? 1 : 0 });\n          if (answered === qEls.length) { finish(container, mode, score, qEls.length, missed); }\n        });\n      });\n    });\n  }\n  var box = qs(root, '.msq-box');\n  setup(root, 'pre');\n  if (box) { var skipFirst = box.hasAttribute('data-default-open'); box.addEventListener('toggle', function(){ if (!box.open) { return; } if (skipFirst) { skipFirst = false; return; } track('quiz_open', { mode: 'pre' }); }); }\n  \/* The end block sits far below this script, so it does not exist yet while this runs\n     during parsing. A delegated listener finds it at click time. *\/\n  document.addEventListener('click', function(e){\n    var startBtn = (e.target && e.target.closest) ? e.target.closest('.msq-start-test') : null;\n    if (!startBtn) { return; }\n    var end = startBtn.closest('.msq-end');\n    if (!end) { return; }\n    var mount = qs(end, '.msq-mount');\n    var clone = qs(root, '.msq-body').cloneNode(true);\n    var intro = qs(clone, '.msq-intro');\n    if (intro) { intro.textContent = 'The same five questions, now that you have read the page.'; }\n    var res = qs(clone, '.msq-result'); if (res) { res.hidden = true; res.innerHTML = ''; }\n    mount.innerHTML = ''; mount.appendChild(clone);\n    setup(mount, 'test');\n    startBtn.hidden = true;\n    track('quiz_open', { mode: 'test' });\n    mount.scrollIntoView({ behavior: 'smooth', block: 'start' });\n    var first = qs(mount, '.msq-opt'); if (first) { first.focus({ preventScroll: true }); }\n  });\n  \/* Shared tool behaviour: checklists remember their ticks on this device, and a chooser\n     shows the result for the option picked. Calculators carry their own small script. *\/\n  function initTools(){\n    qsa(document, '.msq-check').forEach(function(box){\n      var key = 'msq-check-' + meta.slug + '-' + (box.getAttribute('data-id') || 'x');\n      var boxes = qsa(box, 'input[type=checkbox]'), count = qs(box, '.msq-check-count');\n      var yes = qs(box, '.msq-verdict-yes'), no = qs(box, '.msq-verdict-no');\n      var saved = {};\n      try { saved = JSON.parse(localStorage.getItem(key) || '{}'); } catch (e) {}\n      boxes.forEach(function(cb, i){ if (saved[i]) { cb.checked = true; } });\n      function refresh(){\n        var n = boxes.filter(function(cb){ return cb.checked; }).length, state = {};\n        boxes.forEach(function(cb, i){ if (cb.checked) { state[i] = 1; } });\n        try { localStorage.setItem(key, JSON.stringify(state)); } catch (e) {}\n        if (count) { count.textContent = n + ' of ' + boxes.length + (box.getAttribute('data-verb') || ' done'); }\n        if (yes && no) { yes.hidden = n === 0; no.hidden = n !== 0; }\n      }\n      boxes.forEach(function(cb){ cb.addEventListener('change', function(){ refresh(); track('tool_check', { tool: box.getAttribute('data-id') }); }); });\n      refresh();\n    });\n    qsa(document, '.msq-choose').forEach(function(box){\n      var results = qsa(box, '.msq-choice-result');\n      qsa(box, 'input[type=radio]').forEach(function(r){\n        r.addEventListener('change', function(){\n          results.forEach(function(el){ el.hidden = el.getAttribute('data-for') !== r.value; });\n          track('tool_choose', { tool: box.getAttribute('data-id'), choice: r.value });\n        });\n      });\n    });\n  }\n  \/\/ read progress: how far down the article the reader has scrolled, fired once per quarter\n  (function(){\n    var art = root.closest('article') || root.parentNode, fired = {};\n    function check(){\n      var top = art.getBoundingClientRect().top + window.pageYOffset, h = art.offsetHeight || 1;\n      var seen = (window.pageYOffset + window.innerHeight - top) \/ h * 100;\n      [25, 50, 75, 100].forEach(function(m){ if (seen >= m && !fired[m]) { fired[m] = true; track('read_progress', { percent: m }); } });\n    }\n    window.addEventListener('scroll', check, { passive: true }); window.addEventListener('load', check);\n  })();\n  if (document.readyState === 'loading') { document.addEventListener('DOMContentLoaded', initTools); } else { initTools(); }\n})();<\/script>\n<\/div>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"how-many-ai-researcher-jobs-do-the-labs-actually-list\">How many AI researcher jobs do the labs actually list?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The pages that rank for this search are career guides with no listings in them, so this page starts with the listings. On 23 October 2026 the <a href=\"https:\/\/jobs.ashbyhq.com\/openai\" target=\"_blank\" rel=\"noopener\">OpenAI board<\/a> carried 802 postings, the <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\" target=\"_blank\" rel=\"noopener\">Anthropic board<\/a> 682, <a href=\"https:\/\/www.google.com\/about\/careers\/applications\/jobs\/results?company=DeepMind\" target=\"_blank\" rel=\"noopener\">Google Careers filtered to DeepMind<\/a> 120, the <a href=\"https:\/\/job-boards.greenhouse.io\/xai\" target=\"_blank\" rel=\"noopener\">xAI board<\/a> 298 and the <a href=\"https:\/\/jobs.ashbyhq.com\/cohere\" target=\"_blank\" rel=\"noopener\">Cohere board<\/a> 114. A posting counts here as a research listing when its title says research scientist, research engineer or researcher; program managers, communications leads, economists, analysts and people researchers are left out.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That rule gives 124 research listings across the four labs, plus four at Cohere, and one oddity: xAI lists zero. Its board titles that work Member of Technical Staff, eight postings on the day, so a search for researcher at xAI finds nothing and a reader who stops there gets the wrong answer.<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Board, read 23 October 2026<\/th><th>All postings<\/th><th>Research titled<\/th><th>Share of board<\/th><th>Titles the rule caught<\/th><\/tr><\/thead><tbody><tr><td>OpenAI (Ashby)<\/td><td>802<\/td><td>35<\/td><td>4.4%<\/td><td>Researcher, Research Engineer, Research Scientist<\/td><\/tr><tr><td>Anthropic (Greenhouse)<\/td><td>682<\/td><td>42<\/td><td>6.2%<\/td><td>Research Engineer (most), Research Scientist, Staff Research Engineer<\/td><\/tr><tr><td>Google DeepMind (Google Careers)<\/td><td>120<\/td><td>47<\/td><td>39.2%<\/td><td>Research Scientist 23, Research Engineer 22, both in one title 2<\/td><\/tr><tr><td>xAI (Greenhouse)<\/td><td>298<\/td><td>0<\/td><td>0%<\/td><td>None; eight Member of Technical Staff postings instead<\/td><\/tr><tr><td>Cohere (Ashby)<\/td><td>114<\/td><td>4<\/td><td>3.5%<\/td><td>Senior Research Scientist, Staff and Senior Research Engineer<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Research is a thin slice of every board except DeepMind&#8217;s, and it is not remote. At OpenAI 34 of the 35 sit in San Francisco and one in London. Anthropic spreads its 42 across San Francisco, New York, Seattle, London and Zurich, with three tagged remote friendly and travel required. DeepMind&#8217;s 47 split 28 in the United States, 16 in Britain, two in Japan and one each in Switzerland and India. The counts of <a href=\"https:\/\/getaiwork.com\/articles\/anthropic-jobs\/\">every Anthropic posting<\/a> and <a href=\"https:\/\/getaiwork.com\/articles\/openai-careers\/\">every OpenAI role<\/a> on this site go through the whole boards by department. Meta, which is missing from this count, shows the same thinness: <a href=\"https:\/\/getaiwork.com\/articles\/meta-ai-jobs\/\">Meta&#8217;s own board on 24 October 2026<\/a> put 98 of its 1,280 postings under its Artificial Intelligence team filter.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1376\" height=\"768\" src=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/editing-resume-for-open-ai-researcher-jobs.jpg\" alt=\"A close up of hands editing a resume to apply for ai researcher jobs.\" class=\"wp-image-1219\" srcset=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/editing-resume-for-open-ai-researcher-jobs.jpg 1376w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/editing-resume-for-open-ai-researcher-jobs-300x167.jpg 300w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/editing-resume-for-open-ai-researcher-jobs-1024x572.jpg 1024w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/editing-resume-for-open-ai-researcher-jobs-768x429.jpg 768w\" sizes=\"auto, (max-width: 1376px) 100vw, 1376px\" \/><figcaption class=\"wp-block-image__caption\">Many open listings do not strictly require an advanced academic degree.<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"do-ai-researcher-jobs-require-a-phd\">Do AI researcher jobs require a PhD?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Read the listings and the honest answer is: rarely in writing, often in practice. Across the 124 research listings, 85 never mention a doctorate at all. 36 name a PhD and in the same breath offer a way round it, usually the phrase or equivalent practical experience. Only three state a PhD under their qualifications and nothing else: one at DeepMind and two at Anthropic. That is 2.4% of the research listings on 23 October 2026.<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Lab, 23 October 2026<\/th><th>Research listings<\/th><th>No PhD wording<\/th><th>PhD named, with an alternative<\/th><th>PhD stated, no alternative read<\/th><\/tr><\/thead><tbody><tr><td>OpenAI<\/td><td>35<\/td><td>29<\/td><td>6<\/td><td>0<\/td><\/tr><tr><td>Anthropic<\/td><td>42<\/td><td>36<\/td><td>4<\/td><td>2<\/td><\/tr><tr><td>Google DeepMind<\/td><td>47<\/td><td>20<\/td><td>26<\/td><td>1<\/td><\/tr><tr><td>xAI<\/td><td>0<\/td><td>0<\/td><td>0<\/td><td>0<\/td><\/tr><tr><td>Four labs together<\/td><td>124<\/td><td>85<\/td><td>36<\/td><td>3<\/td><\/tr><tr><td>Cohere, for comparison<\/td><td>4<\/td><td>3<\/td><td>0<\/td><td>1<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">The wording differs by lab. OpenAI&#8217;s six PhD sentences all leave a door open: the <a href=\"https:\/\/jobs.ashbyhq.com\/openai\/9b11373c-1643-4ea6-bbcd-033d5b8a0d3e\" target=\"_blank\" rel=\"noopener\">Researcher, Safety Oversight posting<\/a> asks you to hold a Ph.D. or other degree in computer science, machine learning or a related field, and the <a href=\"https:\/\/jobs.ashbyhq.com\/openai\/941bad28-7abe-43c7-b20a-2bc7e5b3c6e8\" target=\"_blank\" rel=\"noopener\">Researcher, Alignment posting<\/a> asks for a PhD or equivalent experience in research. DeepMind writes one clause 24 times, PhD degree or equivalent practical experience, and its single hard case, the <a href=\"https:\/\/www.google.com\/about\/careers\/applications\/jobs\/results\/101711965469975238-research-scientist-humanai-interaction-deepmind\" target=\"_blank\" rel=\"noopener\">Research Scientist, Human AI Interaction<\/a> role, wants a PhD plus six years of post PhD research, a senior academic post wearing a job title.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Anthropic is the strange one. All 42 research listings carry the same Logistics line, minimum education a bachelor&#8217;s degree or an equivalent combination of education, training and experience, and its <a href=\"https:\/\/www.anthropic.com\/careers\" target=\"_blank\" rel=\"noopener\">careers page<\/a> says about half of its technical staff have PhDs but plenty of brilliant colleagues never went to college. Yet two life sciences postings, including <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/5285250008\" target=\"_blank\" rel=\"noopener\">Research Scientist, Life Sciences (Experimental Biology)<\/a>, list a Ph.D. as a minimum qualification higher up the same page. Both lines are on the posting; this page reports the contradiction rather than picking a side.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1600\" height=\"900\" src=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-chart.png\" alt=\"Bar chart of PhD wording across 124 ai researcher jobs listings on 23 October 2026\" class=\"wp-image-1223\" srcset=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-chart.png 1600w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-chart-300x169.png 300w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-chart-1024x576.png 1024w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-chart-768x432.png 768w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/ai-researcher-jobs-chart-1536x864.png 1536w\" sizes=\"auto, (max-width: 1600px) 100vw, 1600px\" \/><figcaption class=\"wp-block-image__caption\">Of 124 research titled listings on the OpenAI, Anthropic and Google DeepMind boards on 23 October 2026, 85 never mention a PhD, 36 name one with an alternative, and 3 state one with no alternative. xAI listed no research titled roles.<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"what-do-the-listings-ask-for-instead-of-a-phd\">What do the listings ask for instead of a PhD?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The 85 listings with no PhD wording are not blank. They ask for the evidence a doctorate usually produces, without naming the degree. OpenAI&#8217;s plain <a href=\"https:\/\/jobs.ashbyhq.com\/openai\/5f0c6579-0bfb-4a06-8a43-1dd371499e10\" target=\"_blank\" rel=\"noopener\">Research Scientist posting<\/a> wants a track record of coming up with new ideas, demonstrated by accomplishments such as first author publications or projects. Its newest research listing on the day, <a href=\"https:\/\/jobs.ashbyhq.com\/openai\/c955617d-f465-4f7a-b57c-98284fd14777\" target=\"_blank\" rel=\"noopener\">Researcher, Misalignment Monitoring Lead<\/a> posted on 22 October 2026, asks for a track record of technical research leadership. Track record appears in nine of OpenAI&#8217;s 35 research listings; PhD appears in six.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Anthropic goes further than silence. Five of its research listings, including <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/4613568008\" target=\"_blank\" rel=\"noopener\">Research Engineer, Machine Learning (Reinforcement Learning)<\/a>, print a list headed strong candidates need not have, and its first two items are formal certifications or education credentials and academic research experience or publication history. The certificate advert promising an AI research career for the price of a course has no counterpart on any lab board: no posting read on 23 October 2026 asked for one.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">DeepMind&#8217;s 20 non PhD research listings mostly open with a bachelor&#8217;s degree or equivalent practical experience and then count years: two years with JAX or PyTorch style libraries for the <a href=\"https:\/\/www.google.com\/about\/careers\/applications\/jobs\/results\/142611399284007622-research-engineer-gemini-large-scale-pretraining-deepmind\" target=\"_blank\" rel=\"noopener\">Research Engineer, Gemini Large Scale Pretraining<\/a> role, five to eight years of software development for the staff level ones. One, <a href=\"https:\/\/www.google.com\/about\/careers\/applications\/jobs\/results\/114077679900074694-research-engineer-polaris-deepmind\" target=\"_blank\" rel=\"noopener\">Research Engineer, Polaris<\/a>, lists two minimum qualifications and the first is experience using AI coding agents. The pattern across all three boards is publications or shipped systems, named frameworks and years, with the degree as one route to those and not the only one.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">If you are weighing which of these routes fits your own background, the <a href=\"https:\/\/getaiwork.com\/?h=jobs&#038;utm_source=articles&#038;utm_medium=organic&#038;utm_content=ai-researcher-jobs\">two minute match quiz<\/a> sorts the free board&#8217;s screened listings by what you already have rather than what a course wants to sell you.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"what-do-ai-researcher-jobs-pay-as-listed\">What do AI researcher jobs pay, as listed?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Pay here is quoted only as printed on the posting on 23 October 2026. Listed bands change, offers differ from bands, and nothing on this page is a promise of anything. 34 of OpenAI&#8217;s 35 research listings print a US base salary band, the lowest floor $250,000 and the highest ceiling $850,000, each adding that it offers equity; the London Researcher, Training role prints pounds. Anthropic prints a band on 40 of its 42: 34 in dollars from $300,000 to $850,000, the <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/5459098008\" target=\"_blank\" rel=\"noopener\">Research Engineer, Chip Design RL<\/a> posting of 21 October 2026 at $500,000 to $850,000, and six in pounds for London, the <a href=\"https:\/\/job-boards.greenhouse.io\/anthropic\/jobs\/5119713008\" target=\"_blank\" rel=\"noopener\">Pretraining<\/a> role at \u00a3260,000 to \u00a3630,000. xAI&#8217;s <a href=\"https:\/\/job-boards.greenhouse.io\/xai\/jobs\/5086324007\" target=\"_blank\" rel=\"noopener\">Member of Technical Staff, Model Training<\/a> prints $180,000 to $600,000. Google Careers prints a US base salary range on each DeepMind page, which this count did not extract.<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Board, as listed 23 October 2026<\/th><th>Research listings printing a band<\/th><th>Lowest floor<\/th><th>Highest ceiling<\/th><th>Currency<\/th><\/tr><\/thead><tbody><tr><td>OpenAI<\/td><td>34 of 35<\/td><td>$250,000<\/td><td>$850,000<\/td><td>USD, one GBP<\/td><\/tr><tr><td>Anthropic<\/td><td>40 of 42<\/td><td>$300,000<\/td><td>$850,000<\/td><td>USD 34, GBP 6<\/td><\/tr><tr><td>xAI, Member of Technical Staff<\/td><td>7 of 8<\/td><td>$180,000<\/td><td>$600,000<\/td><td>USD<\/td><\/tr><tr><td>Google DeepMind<\/td><td>printed per page<\/td><td>not extracted<\/td><td>not extracted<\/td><td>USD<\/td><\/tr><tr><td>getAIwork board, STEM background, <a href=\"https:\/\/getaiwork.com\/stats\">stats<\/a> 22 October 2026<\/td><td>173 listings<\/td><td>$156,000 at the 25th percentile<\/td><td>$249,600 at the 75th<\/td><td>USD annual, median $176,800<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">For scale, the getAIwork board&#8217;s own data on 22 October 2026 put the median listed annual pay for roles open to a STEM background at $176,800 across 173 listings, with 1,072 of its 1,716 live listings tagged advanced. The lab research bands sit far above that median and far above almost everything else in AI work, which is exactly why <a href=\"https:\/\/getaiwork.com\/articles\/highest-paying-ai-jobs\/\">the highest paying AI jobs<\/a> on this site are read from listings and not from adverts.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1376\" height=\"768\" src=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/man-with-headphones-applying-for-ai-researcher-jobs.jpg\" alt=\"A man with headphones planning his application strategy for ai researcher jobs.\" class=\"wp-image-1220\" srcset=\"https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/man-with-headphones-applying-for-ai-researcher-jobs.jpg 1376w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/man-with-headphones-applying-for-ai-researcher-jobs-300x167.jpg 300w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/man-with-headphones-applying-for-ai-researcher-jobs-1024x572.jpg 1024w, https:\/\/getaiwork.com\/articles\/wp-content\/uploads\/2026\/10\/man-with-headphones-applying-for-ai-researcher-jobs-768x429.jpg 768w\" sizes=\"auto, (max-width: 1376px) 100vw, 1376px\" \/><figcaption class=\"wp-block-image__caption\">Accounts from two 2026 forum threads sit beside the listings, labelled as what people report.<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"what-do-people-who-got-research-roles-without-a-phd-report\">What do people who got research roles without a PhD report?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Two r\/MachineLearning threads from 2026 carry the accounts the listings cannot. In a <a href=\"https:\/\/www.reddit.com\/r\/MachineLearning\/comments\/1r0xtzd\/d_for_those_of_you_who_secured_research_scientist\/\" target=\"_blank\" rel=\"noopener\">February 2026 thread<\/a> asking people who landed research scientist roles at large labs what their profile looked like, 50 comments were read for this page. One commenter, with no PhD, describes an hour&#8217;s conversation at NeurIPS with someone who turned out to be a research director and being fast tracked into interviews. Another was recruited into Amazon as a research scientist with a bachelor&#8217;s degree, from a tiny startup, on the strength of client work, and adds that an online application today would not get a look. A third, who interviews research scientists, writes that NeurIPS papers get the interview and fundamentals decide it. A postdoc says every friend now in research at Google, NVIDIA or Anthropic had a prior connection to the team.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The <a href=\"https:\/\/www.reddit.com\/r\/MachineLearning\/comments\/1sptj32\/advice_on_becoming_a_research_engineer_d\/\" target=\"_blank\" rel=\"noopener\">April 2026 thread<\/a> is a staff software engineer over 40 asking how to become a research engineer, 56 comments read. Several replies say a publication history is needed and a PhD is simply the easiest machine for producing one. A former research engineer says the common ask is a research master&#8217;s plus work experience or a PhD. One commenter reports getting a research engineer role that month without a PhD, with a research master&#8217;s, roughly eight years of experience and papers in industry tracks. One dissents that the thread is wrong about publications, without saying why.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Set against the listings, the accounts agree that the or equivalent experience clause is real, and that the people who used it had a publication record or a person inside. The postings read as open; the people who got in describe a door that opens from the inside. This is what people report, not a measurement, and the sample is two threads.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"how-do-you-tell-a-real-ai-researcher-job-from-the-noise\">How do you tell a real AI researcher job from the noise?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Three tests, which the 124 listings pass and most adverts fail. The posting sits on the lab&#8217;s own board, at the Ashby, Greenhouse or Google Careers address, not on a Telegram channel claiming to hire for OpenAI. It prints a pay band in a named currency. And its qualification block names evidence, publications, frameworks, years, rather than a certificate it also happens to sell. A role that quotes a rate nobody is paid, or asks for a fee, is not a research job whatever its title says.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The <a href=\"https:\/\/getaiwork.com\/board\/\">free public board<\/a> at getAIwork screens for those tests before a listing goes up, and if you would rather be matched than scroll, the <a href=\"https:\/\/getaiwork.com\/?h=jobsnoexp&#038;utm_source=articles&#038;utm_medium=organic&#038;utm_content=ai-researcher-jobs\">match quiz for readers without lab experience<\/a> takes two minutes. Lab counts move weekly: Anthropic went from 623 postings on 25 September to 682 on 23 October 2026, and DeepMind from 98 on 9 October to 120, so treat every number here as a dated reading and the <a href=\"https:\/\/getaiwork.com\/articles\/google-deepmind-jobs\/\">DeepMind count by job family<\/a> on this site as the earlier snapshot.<\/p>\n\n\n\n<div class=\"msq-tool msq-check\" data-id=\"research-listing-read\" data-verb=\" found\">\n<h3>Read a research listing the way this count did<\/h3>\n<p>Six lines to find on any lab posting before you spend an evening on the application. They restate what the tables above say; they promise nothing.<\/p>\n<label><input type=\"checkbox\">The posting sits on the lab&#8217;s own board: an Ashby, Greenhouse or Google Careers address, not a recruiter&#8217;s channel<\/label>\n<label><input type=\"checkbox\">The title says research scientist, research engineer or researcher, or at xAI, Member of Technical Staff<\/label>\n<label><input type=\"checkbox\">The PhD sentence, if there is one, and whether it continues with or equivalent experience, or other degree, or a master&#8217;s<\/label>\n<label><input type=\"checkbox\">Any Logistics or minimum education line lower on the page, which at Anthropic says a bachelor&#8217;s degree or an equivalent combination<\/label>\n<label><input type=\"checkbox\">What the qualification block asks for instead: publications, a track record, named frameworks, years<\/label>\n<label><input type=\"checkbox\">A pay band in a named currency, quoted as listed on the day, and the location tag<\/label>\n<p class=\"msq-check-count\"><\/p>\n<p class=\"msq-note\">Listings change weekly. The counts on this page were read on 23 October 2026 and nothing here is a promise of an interview or an offer.<\/p>\n<\/div>\n\n\n\n<h2 class=\"wp-block-heading\" id=\"frequently-asked-questions\">Frequently asked questions<\/h2>\n\n\n<div id=\"rank-math-faq\" class=\"rank-math-block\">\n<div class=\"rank-math-list \">\n<div id=\"faq-q-1\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">How many AI researcher jobs are open at OpenAI right now?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>35 of the 802 postings on OpenAI&#8217;s own Ashby board carried researcher, research scientist or research engineer in the title on 23 October 2026, 34 of them in San Francisco and one in London. The count changes weekly.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-q-2\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">Does Google DeepMind require a PhD for research roles?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>On 23 October 2026, 26 of DeepMind&#8217;s 47 research titled listings named a PhD and all but one added or equivalent practical experience or accepted a master&#8217;s. One role asked for a PhD with six years of post PhD research and no alternative.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-q-3\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">Why does xAI show no researcher jobs?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>xAI&#8217;s Greenhouse board titles that work Member of Technical Staff, eight postings on 23 October 2026, none of which mention a degree. A search for researcher on its board returns only a Research Analyst role in another department.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-q-4\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">Is a research engineer the same as a research scientist?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>The labs blur it: 17 Anthropic and OpenAI listings on 23 October 2026 carry both titles with a slash. Reddit accounts from 2026 describe research engineers as supporting experiments and infrastructure while scientists drive the questions, with overlap.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-q-5\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">Can an AI certificate get you an AI researcher job?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>None of the 124 lab research listings read on 23 October 2026 asked for a certificate. Five Anthropic postings say strong candidates need not have formal certifications or education credentials at all.<\/p>\n\n<\/div>\n<\/div>\n<div id=\"faq-q-6\" class=\"rank-math-list-item\">\n<h3 class=\"rank-math-question \">Where can I see AI researcher listings that are not at the big labs?<\/h3>\n<div class=\"rank-math-answer \">\n\n<p>The free public board at getaiwork.com\/board\/ lists screened roles with pay quoted only as the posting states it; on 22 October 2026 it held 1,716 live listings, 1,072 of them tagged advanced.<\/p>\n\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n\n\n<div class=\"msq-end\" id=\"ms-quiz-end\">\n<p><strong>Finished the page?<\/strong> The same five questions from the top, to see what stuck.<\/p>\n<button type=\"button\" class=\"msq-btn msq-start-test\">Test Your Understanding of This Article<\/button>\n<div class=\"msq-mount\"><\/div>\n<\/div>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Platform statuses and listed pay change without notice. Pay figures are quoted only as listed by the poster on the date shown and are never a promise of earnings. This page was verified on the date shown and is rechecked on the weekly refresh.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Every research titled listing on the OpenAI, Anthropic, Google DeepMind and xAI boards counted on 23 October 2026, with the PhD line read on each one and pay only as listed.<\/p>\n","protected":false},"author":2,"featured_media":1218,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","rank_math_title":"AI Researcher Jobs: What 124 Lab Listings Ask For, Counted","rank_math_description":"AI researcher jobs counted on the labs' own boards on 23 October 2026: 124 listings, 39 naming a PhD and only three stating one with no alternative.","rank_math_focus_keyword":"ai researcher jobs","rank_math_canonical_url":"","rank_math_breadcrumb_title":"","rank_math_facebook_title":"","rank_math_facebook_description":"","rank_math_facebook_image":"","rank_math_twitter_title":"","rank_math_twitter_description":"","rank_math_twitter_image":"","rank_math_pillar_content":"","rank_math_robots":[],"rank_math_advanced_robots":[]},"categories":[3],"tags":[],"class_list":["post-1221","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-jobs"],"_links":{"self":[{"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/posts\/1221","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/comments?post=1221"}],"version-history":[{"count":5,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/posts\/1221\/revisions"}],"predecessor-version":[{"id":1245,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/posts\/1221\/revisions\/1245"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/media\/1218"}],"wp:attachment":[{"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/media?parent=1221"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/categories?post=1221"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/getaiwork.com\/articles\/wp-json\/wp\/v2\/tags?post=1221"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}