631 roles openBrowse roles
Best for
DataAnnotation. Careful generalists and coders with no specialist credential. Screening: unpaid written assessment, 1–3 hours, usually one attempt only.
The catch: A long unpaid assessment with a low acceptance rate and no retakes.
Best for
Braintrust. Experts who want task-based AI work rather than waiting on a project match. Screening: profile, then an AI skills interview, then you browse and accept tasks.
The catch: The AI-training work sits inside a general freelance marketplace, so expert tasks share the site with ordinary contract listings, and the bottom of the pay band is annotation work rather than expert work.
Best for
micro1. Software engineers, especially with competitive-programming experience. Screening: aI interview plus a technical assessment.
The catch: Engineering-focused, so there is little here for non-technical fields.
Best for
Mercor. Credentialed experts — MDs, JDs, PhDs, senior engineers, finance professionals. Screening: recorded AI interview, 15–30 minutes, after a CV upload.
The catch: The hardest screen of any of them, and being accepted does not mean work is waiting.
Best for
Handshake AI. PhDs, graduate students and credentialed academics. Screening: application with credential verification.
The catch: A newer expert marketplace, so project volume is still uneven.
Some links here are referral links, which pay this site if you are hired. The order above is by fit for each background, not by which ones pay us.
From the live board
Before choosing a platform, check whether your field has open work. These are today's open roles by field across the platforms we track.
Or read the guide for your profession:
FAQ
It depends on your background, not on the platform. DataAnnotation suits careful generalists in the US, UK, Canada, Australia or New Zealand. Braintrust suits people who want task-based work without a long unpaid assessment. micro1 suits software engineers who want engineering work rather than annotation. Mercor suits anyone with a degree, licence or senior professional experience. Handshake AI suits graduate students and PhDs.
Yes. Most experienced contributors run two or three platforms, because acceptance on one does not guarantee work. The only limit is a conflict of interest on a specific project, which the platform will flag.
Platforms rarely say. The most common reasons are a rushed assessment, a CV that does not make the specialty obvious, or a country the project does not accept. Rejection on one platform says nothing about another.
Get new roles by email
One short email when something opens in your field. No digest spam, and you can unsubscribe in one click.