🛰️ AI Eval Job Radar

Where to actually get AI evaluation work. Pay and fit notes are hand-curated and honest. Link and signup status refresh weekly. This is a read-only snapshot.

Status last refreshed: not yet refreshed · source on GitHub

Tier 1 - best fit

PlatformSignupPay notesWhy it fits
⚪ DataAnnotation.tech not checked Reported as one of the better crowd-AI payers, but rate varies by project. Verify on offer. LLM training, coding, and reasoning tasks. Coding background helps. Entry is a paid trial.
⚪ Mercor not checked Project and seniority dependent. Verify on offer. Matches experts to AI-lab projects via interview. A strong GitHub profile is a real advantage here.
⚪ Mindrift (Toloka) not checked Per-project rates. Comes in waves. Verify on offer. Freelance AI-tutor projects for domain experts including coding and analytics. Flexible.
⚪ Surge AI not checked Reputable payer. Selective intake. Verify on offer. High-quality RLHF and evaluation data. Selective, good reputation, rubric-style work.
⚪ Turing not checked Varies by role and region. Verify on offer. AGI/LLM training plus coding expert work. Dev-leaning, so your repos are an asset.
⚪ Kaggle not checked Prize-based, not wages. Some comps have large cash pools; many are for medals and ranking. ML and data-science competitions. Strong fit for your analytics background. Open to anyone.
⚪ DrivenData not checked Prize-based. Cash prizes on many competitions. Data-science-for-good competitions. Clear problems, open entry, good fit.
⚪ Devpost not checked Prize-based, often cash plus cloud credits. Varies per hackathon. Hackathon hub. AI-agent and DevOps themed events run constantly; your gatekeeper and Compass work fits well.
⚪ RWS / Article One Partners not checked Per-study rewards / bounties for accepted prior art. Not hourly. Patent prior-art research (Article One Partners, now RWS). You actively submit studies here. This is NOT the RWS/Appen crowd AI-eval that was excluded.

Tier 2 - solid

PlatformSignupPay notesWhy it fits
⚪ Toloka AI not checked Per-task. Expert tracks pay more than legacy microtasks. Verify on offer. Older crowdsourcing platform that pivoted into LLM eval and expert tasks.
⚪ Invisible Technologies not checked Decent by reputation. Role dependent. Verify on offer. AI training and LLM QA work for labs. More ops-flavored.
⚪ Defined.ai not checked Project dependent. Verify on offer. Speech, text, and LLM data marketplace. Corporate clients, slower onboarding.
⚪ iMerit (Ango Hub) not checked Enterprise contracts. Verify on offer. Annotation plus expert tasks. More structured and enterprise.
⚪ Innodata not checked Enterprise. Verify on offer. Enterprise AI data prep with expert annotator programs.
⚪ Sapien AI not checked Newer, rates unproven. Verify on offer. Crowd labeling and RLHF, newer player.
⚪ Pareto.ai not checked Startup, rates vary. Verify on offer. Data-labeling startup doing expert and RLHF tasks.
⚪ Micro1 not checked Placement dependent. Verify on offer. AI-driven vetting and recruiting that places experts into AI data work.
⚪ AIcrowd not checked Prize-based. Varies by challenge. AI and ML research challenges. More research-flavored.
⚪ Zindi not checked Prize-based. Cash prizes on many. AI competitions, Africa-focused but open globally. Real-world datasets.
⚪ HackerEarth not checked Prize-based; some challenges are hiring-linked. Hackathons and hiring challenges. Mixed dev and ML.
⚪ Topcoder not checked Pays per challenge. Competitive fields. Algorithm, development, and design challenges that pay out. More traditional dev.

Tier 3 - long shots

PlatformSignupPay notesWhy it fits
⚪ Clickworker / UHRS not checked Low, per-task. Filler work. Microtask crowd, some AI tasks. Filler-tier pay but always-on.
⚪ Anthropic not checked Strong when a role exists, but selective and intermittent. A company careers page, not an open application. Safety/red-team/contractor data roles show up only now and then. Watch, do not expect a door that is always open.
⚪ OpenAI not checked High when active, very selective. The old /contributions link is unreliable; use careers. Also a company careers page, not an open application. Runs expert AI-training and contractor programs now and then; search "expert" or "AI trainer". Very selective.
⚪ Prolific not checked Per-study rates set by researchers. Easy entry, uneven volume. Research-study platform. Hosts AI/LLM behavior experiments beyond surveys. Easy to start.
⚪ Humanloop not checked Not gig work. Salaried roles only. NOTE - this is an LLM eval/dev tooling company, not a gig platform. Watch for full-time/contract roles.
⚪ Major League Hacking (MLH) not checked Mostly prizes and swag, not cash-heavy. Open hackathon season. Student-leaning, but anyone can join member events.
⚪ ARC Prize not checked Prize-based, not wages. Big pool, very hard. Reasoning benchmark with a large prize pool. Hard, but squarely in your wheelhouse.