Handshake AI Review 2026: Who Qualifies and What It Lists

Handshake AI Review 2026: Who Qualifies and What It Lists

By Mufy Pachorawala, founder of getAIwork · Last updated: September 3, 2026

Handshake AI is a paid fellowship programme that recruits people to help train and evaluate AI models. It is open to US-based applicants with valid work authorisation, from associate degree through PhD and postdoc, and states that no AI experience is required. Listed rates on its own site run from $40 an hour for AI evaluation up to $125 an hour and above for specialist roles. Handshake reports more than 750,000 fellows and $100 million paid out in total.

Key facts

  • Run by Handshake, the university careers platform, which gives it an unusual recruiting channel straight into student and graduate populations
  • Company-stated scale: 750,000+ fellows and $100 million in total payouts
  • Listed rates by role (as listed on Handshake’s own site, not a promise of earnings): AI Evaluation Specialist $40/hr, Software Engineer $65/hr, Energy Professional $80/hr, FP&A Analyst $80/hr, Philosophy Expert $120/hr, Game Developer $125/hr, PCB Tool Specialist $125/hr
  • Eligibility is the real filter: US-based with valid work authorisation. F-1 students on CPT or OPT are eligible; STEM OPT is excluded
  • Five stated steps: apply and verify identity, get matched to a project, complete training, contribute, get paid
  • 591 AI platform programmes are live on the free getAIwork board as of September 3, 2026

What is Handshake AI?

Handshake AI is the AI-data arm of Handshake, the careers platform that most American universities plug their students into. It pays people to help train and evaluate frontier AI models, and it calls its contributors fellows rather than annotators or contractors, which tells you something about the audience it is aiming at.

The strategic logic is neat enough that it is worth spelling out. Every AI lab needs credentialled humans. Handshake already had a verified pipeline of students, graduates and alumni sorted by degree, university and field. Turning that into an expert-supply business is a shorter step than building an expert network from scratch, which is what most competitors in this category had to do.

Student working at a laptop, the audience Handshake AI recruits from

In practice that means Handshake AI skews younger and more academic than platforms recruiting from the general remote-work population. Its own materials describe an intake spanning associate degrees through PhDs and postdocs, and it states that no prior AI experience is required, which is accurate: what it is buying is the subject knowledge, not familiarity with machine learning.

Who is eligible for Handshake AI?

This is where most applications end, so read it before spending an evening on the form. Handshake AI requires participants to be based in the United States with valid work authorisation. International students on F-1 status may participate if they are eligible for CPT or OPT, with STEM OPT specifically excluded.

Situation Eligible, per Handshake’s stated rules
US citizen or permanent resident, any degree level from associate upward Yes
F-1 international student eligible for CPT Yes
F-1 international student on standard OPT Yes
F-1 student on STEM OPT extension No, specifically excluded
Based outside the United States No
No AI or machine learning background Not a barrier, by Handshake’s own wording

If you are outside the US, this one is simply not for you, and no amount of a good CV changes that. The good news is that the eligibility rule is the unusual part rather than the norm: several of the larger expert platforms recruit globally, and our list of AI training platforms with open intake flags which ones do.

What does Handshake AI list as pay?

Handshake publishes hourly rates against specific role types on its own site, which is more transparency than most of this category offers. The figures below are as listed by Handshake. They are not an earnings estimate, they do not account for unpaid training or gaps between projects, and no platform can tell you how many hours it will have.

Role as listed Listed hourly rate What it signals
AI Evaluation Specialist $40 The generalist entry point, and the one with the widest intake
Software Engineer $65 Standard engineering evaluation and code data work
Energy Professional $80 Industry credential, narrow supply
FP&A Analyst $80 Finance judgement, rubric-heavy work
Philosophy Expert $120 Reasoning and argument quality, a genuinely scarce skill in this market
Game Developer or Designer $125 Specialist domain with few qualified evaluators
PCB Tool Specialist $125 Very narrow hardware expertise
Engineer at a computer, one of the specialist roles Handshake AI lists

The shape of that table is the useful finding, not any single number in it. The spread from $40 to $125 tracks scarcity almost perfectly. Nobody is paying a premium for AI knowledge here. They are paying a premium for a background that is hard to hire, and a philosophy graduate outranking a software engineer on this list is the clearest evidence you will find that the old assumptions about which degrees pay do not transfer cleanly into AI data work.

Handshake also states that posted rates do not change mid-project. That is a small promise and a meaningful one, because rate changes partway through a project are a recurring complaint across this whole category.

See which platforms match your background

Handshake AI is one of 591 live platform programmes on the getAIwork board. The 2-minute quiz filters them by your skills, level and location.

Take the 2-minute match quiz →

Free to take · real listings, screened daily · no income promises, ever

How the process works

Handshake describes five steps: apply and verify your identity, get matched to a project, complete training, contribute your expertise, and receive payment. Identity verification up front is worth noting because it is a genuine legitimacy signal. Platforms that pay real money in the US have to know who they are paying, and the ones that skip this step are usually the ones that also skip paying.

Person working at a home desk on a fellowship project

The step that catches people out is the second one. Applying is not joining, and joining is not working. Matching is demand-driven, which means an accepted fellow with an unfashionable specialism can sit without a project for weeks while somebody with a scarcer background starts within days. This is not specific to Handshake, and our article on what silence from these platforms actually means covers the general pattern.

The part nobody puts in the pitch

Three things are worth knowing before you build any plans around this.

The headline numbers are cumulative, and yours will not be. $100 million paid across 750,000 fellows is a real figure describing a large programme. It is not a per-person figure, and dividing one by the other would produce a number that means nothing about your own project. Any site doing that division for you is not being straight with you.

Training time is part of the deal. Step three of the five is training, and in this category training is frequently unpaid or paid at a lower rate than the project work itself. Ask, in writing, before you start. A platform with real volume answers that question directly.

Project work is not employment. Fellowship framing is warmer than contractor framing, but the underlying arrangement is project-based and there is no guarantee of continuing work. In the US that also means the tax treatment is yours to handle, which we cover in our guide to the tax side of AI task work.

Our assessment: a legitimate, unusually transparent platform with published rates and a clear process, and a hard eligibility wall around US work authorisation. Strongest fit for US-based students, graduates and credentialled professionals who want project work rather than a steady schedule. Apply, keep other queues running, and treat the listed rate as the rate for hours you actually get, not for hours you hope to get.

How it compares to the other platforms

Handshake AI The wider expert-platform category
Recruiting channel University and graduate populations via Handshake Open web application, general remote-work audience
Geography United States only, work authorisation required Frequently global, with per-project country rules
Rate transparency Rates published per role on the site Usually disclosed only after acceptance
Entry level Associate degree upward, no AI experience needed Varies widely, some genuinely open, some postgraduate only
Work shape Project-based fellowship, matched to demand Project-based, matched to demand

Read across that table and Handshake AI’s real differentiator is not pay, it is transparency plus a recruiting channel nobody else has. The cost of that channel is the eligibility rule. If you clear it, it belongs on your shortlist. If you do not, there is no workaround worth trying.

Browse the live board free

1,435 live AI listings as of September 3, 2026, screened and human-approved, filled ones deleted daily. No signup.

See the free board →

Free and public · no signup · screened and human-approved

Frequently asked questions

Is Handshake AI legit?

Yes. It is run by Handshake, the established university careers platform, verifies contributor identity before payment, publishes hourly rates by role, and reports $100 million in total payouts across more than 750,000 fellows. It charges applicants nothing.

Who can join Handshake AI?

US-based applicants with valid work authorisation, from associate degree through PhD and postdoc. F-1 students eligible for CPT or standard OPT can participate. STEM OPT is excluded. No AI experience is required.

Can I join Handshake AI from outside the US?

No. Handshake AI states that participants must be US-based with valid work authorisation. Several other expert platforms do recruit internationally, which is the practical route if you are outside the US.

What does Handshake AI pay?

As listed on its own site, hourly rates run from $40 for an AI Evaluation Specialist to $125 and above for specialist roles such as game developer and PCB tool specialist. These are listed rates for project work, not an earnings estimate, and hours are not guaranteed.

Do you need AI experience for Handshake AI?

No. Handshake states plainly that no AI experience is required. What it is buying is subject expertise in your own field, and the evaluation work is taught during the training step.

Is Handshake AI a job or a contract?

It is project-based fellowship work rather than employment. There is no guaranteed schedule, projects are matched to demand, and in the US the tax treatment falls to you as self-employment income.

How long does Handshake AI take to match you to a project?

There is no published timeline, and matching is demand-driven. Fellows with scarce specialisms are matched faster than generalists. Extended silence after acceptance usually means no current project needs your background rather than that you were rejected.

Does Handshake AI change rates mid-project?

Handshake states that posted rates are set upfront and do not change mid-project. That is worth having in writing at the start of any project, on any platform.

Mufy Pachorawala

Mufy Pachorawala · Founder, getAIwork

My AI scans thousands of AI-job posts (40,786 screened so far) and I personally approve every listing before it reaches the board. I write these articles by the same rules: pay quoted only as listed, and no income promises. Read our editorial rules.

Related: AI training platforms hiring now · 12 Outlier alternatives compared · Mercor review · AI training jobs with no experience · AI job market statistics

More from the blog