Handshake AI
Runs the Handshake AI Fellowship for paid AI training work.
Best forPeople already on the Handshake career network
- About $1.15B total gross ARR, with the AI training unit near $1B1
Not rated yet
How we rateJump to a company, a ranking or a tool
Updated October 2026
Mercor, micro1 and Handshake AI are large, real businesses: Mercor reports a run rate above $2B[2], Handshake's AI training unit is near $1B[1] and micro1 reached $500M gross[9]. Outlier is operated by Scale AI[6].
Not scored yet, on purpose. Expert scores start with pay, and we will only score pay from a public index of job listings that we are still building. Until then the order below follows how much of the rubric we can evidence, then the name. Ownership, incidents and referral terms are all here today.
Runs the Handshake AI Fellowship for paid AI training work.
Best forPeople already on the Handshake career network
Not rated yet
How we rateHires experts as contractors for AI lab projects.
Best forProfessionals who want hourly expert contracts with AI labs
Watch out for: Confirmed a cyberattack in March 2026; several class actions followed.3,4
Not rated yet
How we rateRecruits experts and professionals for paid AI training projects.
Best forExperts and professionals open to project-based AI work
Not rated yet
How we ratePays contributors for AI training tasks.
Best forGeneralists and specialists who want task-based AI training work
Watch out for: Scale AI is 49% owned by Meta7
Not rated yet
How we rateHires workers through DataAnnotation.tech, reportedly its subsidiary.
Best forWriters and specialists who apply through DataAnnotation
Not rated yet
How we ratePay level and evidence30%
Work availability20%
Payment reliability20%
Onboarding fairness15%
Reputation and incidents15%
A score is published only when criteria worth at least 35% of the weight are evidenced. The overall score is the weighted mean of the evidenced criteria. Commissions are never an input. Full methodology
FAQ
Each audience has its own published rubric of five weighted criteria, and every sub-score cites the facts behind it. A score appears only when we can evidence at least 35% of the rubric's weight; otherwise the card says there is not enough public data. Commissions never enter the calculation.
DataAnnotation.tech is reported to be a Surge AI subsidiary, including by TIME in 2024, but its own website does not name Surge. We show this as reported rather than confirmed.8
We do not publish a pay ranking yet, because we will only score pay from a public index of job listings and that index is still being built. Until it is published, pay scores on this site show as not enough public data rather than a guess.
Handshake pays 10% of the referred fellow's hourly rate, for up to 40 hours, over 6 months. At $50 an hour that is 0.10 x 50 x 40 = $200 at most per referral.11