A Curated Daily Directory

AITasks.live

A curated directory of AI training, annotation, and evaluation tasks across Mercor, Outlier, DataAnnotation, and every other major platform.

8+
Platforms Tracked
12
Role Types
Daily
Curation Cadence
§ 01 · What This Is

One trustworthy place to find active AI training tasks.

The AI training-data market has fragmented across dozens of platforms — Mercor, Outlier, DataAnnotation, Surge AI, Invisible, Prolific, Scale AI, and a long tail of smaller players. Each has its own application flow, listing format, pay structure, and geographic eligibility rules.

Today, contractors piece together their pipeline by checking multiple sites manually, scrolling Reddit threads, and trading tips in Discord servers. There is no single, trustworthy place to see what is newly available across the ecosystem.

AITasks.live fixes that. Hand-picked, structured, refreshed daily — with pay rate, geographic eligibility, and role type surfaced consistently on every listing.

§ 02 · Who It Is For

Contractors who already do this work.

You have a marketable skill: coding, writing, a STEM background, a language, or a professional credential. You treat AI training work as primary or meaningful side income. You already check three to ten platforms regularly. You care about pay transparency and geographic eligibility — especially if you are outside the United States.

Most existing resources are US-centric and bury international workers under irrelevant listings. AITasks.live treats geographic eligibility as a first-class filter, not an afterthought.

§ 03 · Platforms Tracked

The major players in AI training data work.

Each platform has its own niche — pay tier, task type, eligibility model, and quirks. Editorial notes and live listings appear on the directory at launch.

MercorExpert

Credentialed expert contributor work — doctors, lawyers, engineers, PhDs. Public job board at work.mercor.com/explore. Among the highest pay tiers in the ecosystem.

OutlierAI Training

Scale AI's contributor-facing brand. Coding, writing, STEM, language, and generalist tasks. Login-walled feed; listings are surfaced when verifiable.

DataAnnotationInvite Only

RLHF and evaluation tasks, primarily writing-focused. Application-based entry. Steady reputation in the contractor community.

Surge AIAnnotation

High-quality annotation and RLHF for major model labs. Selective intake. Strong pay relative to general annotation platforms.

InvisibleAgent Ops

AI training and agent operations work. Tasks include prompt engineering, evaluation, and structured business workflows.

ProlificResearch

Academic and applied research studies, including AI evaluation tasks. Lower per-task pay but consistent global eligibility.

Scale AILabeling

The original AI data infrastructure company. Direct Scale work is less accessible to individual contractors; most flow goes through Outlier.

+ Long TailEmerging

Specialty platforms covering language pairs, domain expertise (medicine, law, finance), red-teaming, and emerging model-evaluation work.

§ 04 · What This Work Is Called

The vocabulary varies. The work is the same.

Different platforms use different terminology for overlapping tasks. If you have done any of the work below, AITasks.live is for you.

AI Training Data
The broadest umbrella term — covers any human-generated input used to train, fine-tune, or evaluate AI models.
Data Annotation
The original term, dating to pre-LLM computer vision work. Still widely used. Includes labeling, classification, and bounding-box work.
RLHF
Reinforcement Learning from Human Feedback — comparing two or more model outputs and ranking them for quality, safety, or accuracy. Read the full guide →
Model Evaluation
Sometimes called "evals." Judging individual model responses against quality, factuality, or safety criteria.
Prompt Engineering
Writing prompts designed to test, train, or stress-test models. Often domain-specific.
Expert Contributor
Mercor's framing — credentialed professionals (medicine, law, advanced math, engineering) producing high-value training data.
Red-Teaming
Adversarial probing of model outputs to surface harmful, biased, or unsafe behavior before deployment.
§ 05 · Guides & Comparisons

Reading while you wait.

Working notes from inside the AI-training-data ecosystem — how the work is structured, who pays what, and how to choose between platforms.

Guide
What is RLHF work? A practical guide for contractors
Comparison
Mercor vs Outlier: which platform should you apply to?
§ 06 · Frequently Asked

Common questions about AI training data work.

What is AI training data work?

AI training data work covers a range of contract tasks that help train and evaluate large language models — including data annotation, labeling, RLHF, model evaluation, prompt engineering, and specialty expert contributions. Workers are typically paid hourly or per task.

How much does AI training data work pay?

Pay varies widely by platform, role, and required expertise. General annotation tasks often pay $15–25/hr, while specialized work — coding, medicine, law, advanced mathematics — can pay $40–100+/hr on platforms like Mercor and Outlier.

Is this work available outside the United States?

Yes. Many platforms accept contributors in the EU, UK, India, Latin America, and other regions, though eligibility varies task-by-task. AITasks.live treats geographic eligibility as a first-class filter to make non-US-eligible work easier to find.

How is AITasks.live different from Reddit or Discord?

Reddit and Discord are excellent for community and word-of-mouth, but bad for discovery — listings get buried, eligibility is unclear, and pay information is inconsistent. AITasks.live is editorially curated with structured data: each listing surfaces pay rate, geographic eligibility, role type, and apply link in a consistent format.

Is AITasks.live free?

Yes. The directory is free for contractors. The site is supported by personal referral codes on platform signup links, disclosed transparently.

When does the full directory launch?

Soon. Check back over the coming weeks.