LIVE 2,000+ jobs added today  ·  $20–$200/hr  ·  Apply direct Browse now →
// Frontier models need frontier humans

The hub for LLM trainers & AI experts.

2,000+ pay-disclosed AI training, RLHF, and expert-data roles indexed daily at $20–$200/hr — plus remote freelance and US corporate feeds. Neutral aggregator. Apply direct. Data never resold.

Free forever for candidates  ·  No résumé middleman  ·  Applies open in one click
Freshness · 60s re-poll
Pay disclosed · Every listing
Direct apply · No forms
Anonymous · No login to browse
Never resold · Candidate data
// One feed · Three tracks

Ranked by freshness, pay, and match.

Every listing on HireFeed is pay-disclosed, deduplicated, and dated. Toggle between AI training (default), remote freelance, and US corporate roles — same trust bar, same ranking rules.

Live · Indexed continuously Sorted by freshness × pay × match
RLHF Annotator · LLM Reasoning
Mercor  ·  remote  ·  PhD preferred
$45/hr
2m ago
Coding Evaluator · Python + Debugging
Outlier  ·  remote  ·  SxS eval
$60/hr
4m ago
Domain Expert · Law (US JD)
Mercor  ·  contract reasoning  ·  bar-passed
$120/hr
6m ago
Math Reasoning Expert · PhD
Outlier  ·  proof verification
$80/hr
9m ago
Medical Domain Expert · MD
Surge AI  ·  clinical reasoning
$200/hr
18m ago
Agentic Eval · Tool-use trajectories
Labs direct  ·  adversarial review
$130/hr
27m ago
Senior Software Engineer · Contract
Turing  ·  remote  ·  40h/wk
$150/hr
14m ago
ML Engineer · Contract
Micro1  ·  production LLMs
$220/hr
31m ago
Product Designer · Remote
Toptal  ·  SaaS  ·  30h/wk
$95/hr
42m ago
Technical Writer · API docs
Contra  ·  async  ·  devtools
$70/hr
1h ago
Applied Scientist · Foundation Models
Frontier lab  ·  SF or remote  ·  full-time
$280K–$450K
3h ago
Data Ops Lead · Human evaluation
Enterprise AI co  ·  NYC  ·  hybrid
$190K–$260K
5h ago
RLHF Program Manager
Series-B model lab  ·  remote-US
$170K–$220K
8h ago
// The HireFeed Pay Index

What AI training actually pays.

The only public rate index built from pay-disclosed listings — not surveys. Seven domains, top-decile rates, 90-day trend, published methodology. Neutral.

See the Pay Index
// Medicine · MD
$100–210/hr
Top decile $250/hr  ·  ▲ +12%
// Agentic Eval
$70–130/hr
Top decile $160/hr  ·  ▲ +21%
In build · First cohort Q4 2026
// HireFeed Vetted

Get vetted once. Carry one credential everywhere.

Every AI data platform runs its own screening on the same experts. HireFeed Vetted is a portable credential — candidate-owned, evidence-tiered, designed to be recognized across platforms instead of re-proved per platform.

// Stage 1 · Free

Skill Passport

4 minutes

Declare domains, credentials, languages, coding skills. No résumé upload. Instant personalized feed + realistic pay band estimate.

// Stage 2 · Free

Domain Exam

45–90 minutes · Proctored

Scenario-based, adaptive, co-authored with working RLHF leads. Passing = V1 Verified. Credential claims verified in parallel.

// Stage 3 · Paid

Calibration Trial

2–4 hours · Paid at market rate

Real tasks against a held-out golden set. Measures inter-rater agreement (κ), rubric adherence, throughput. Passing = V2 Calibrated.

// Stage 4 · Ongoing

Sourced & live

Continuous

Vetted experts enter the searchable pool. Post-placement quality reviews update the score continuously. Top decile = V3 Frontier.

// The Skill Passport

Evidence, not résumés.

Résumés don't say whether you can rank two model responses at 2am. HireFeed Vetted profiles carry per-skill evidence tiers — self-report to live production telemetry — so buyers see what you've actually done, not what you've claimed.

  • → Credentials verified via DigiLocker / Persona / registrar APIs
  • → Per-skill scores with visible evidence tier
  • → Rubric-graded samples buyers can inspect
  • → Candidate-owned. Portable. Never resold.
// Skill Passport · Example Tier: coming Q4 2026
Domain Math & Science · PhD Registrar ✓
Skill Rubric writing · 92 Graded sample
Skill Agentic eval · 88 κ=0.84 · Trial
Skill Multi-turn eval · 74 Exam pass
Availability 20h/wk · IST Self · verified
Illustrative example · No live user data shown
// Five defaults every other board gets wrong

The rules the AI training market needs.

// 01

Freshness first

60s re-poll. Closed roles drop within minutes. Queues going dark = an alert, not a surprise.

// 02

Pay disclosed

Every listing shows the rate — hourly, range, or salary. No "competitive comp" hides.

// 03

Direct apply

Click through to the platform's own apply page. No intermediary form. No middleman.

// 04

Anonymous browse

No login required to search. No email wall. No cookie fingerprint sold to recruiters.

// 05

Never resold

Candidate data is not for sale. Not to platforms. Not to advertisers. Full stop.

// For talent teams · Data ops · Frontier labs

Stop screening 10,000 to hire 40.

Every AI data platform runs the same funnel on the same experts. HireFeed Vetted is designed to be the shared credential layer — so your talent team searches a pre-calibrated pool instead of re-running the assessment.

See the sourcing product →
// Search
Query the vetted pool by facet — domain, tier, κ band, language, timezone, rate band. Coming Q4 2026.
// Evidence
Per-skill scores with a visible evidence tier — self-report vs. exam vs. graded sample vs. live production telemetry.
// Neutral
No sponsored rankings, no data resale, no lock-in. The Pay Index and vetting funnel methodologies are public.
// Interop
Portable credential. Recognized across platforms — designed as a shared credential layer for the human-data industry.
// FAQ

Common questions.

What is HireFeed?

A neutral, pay-disclosed job feed for AI training work — RLHF, LLM evaluation, agentic red-teaming, domain-expert data. We index the market so experts can find and apply to real, rate-transparent roles across every major platform in one place. Candidates never pay. Candidate data is never resold.

How much do AI trainers earn in 2026?

Domain-dependent. Generalist annotation pays $18–$25/hr. Coding evaluation $40–$75/hr. Math and science reasoning at PhD level $60–$120/hr. Agentic evaluation and red-teaming $70–$130/hr. Law (JD) $90–$160/hr. Medicine (MD/DO) $100–$210/hr. Top decile in medicine and law reaches $250/hr. See the full Pay Index for methodology.

How is HireFeed different from Outlier, Mercor, or Surge?

Those are marketplaces — they vet talent and sell it at a take rate of 30–55%. HireFeed is an aggregator + credential layer. We do not gate the feed, do not take a cut, do not collect data resale rights, and rank platforms with a publicly disclosed methodology. Our upcoming HireFeed Vetted credential is portable — candidate-owned, not locked to any single marketplace.

Do you charge candidates?

No. Candidates never pay to browse the feed, apply, or hold a HireFeed Vetted credential (when it launches). Revenue comes from buyer-side sourcing seats and programmatic job-ad partners on Tier 3 corporate roles. The candidate side is free forever.

When does the HireFeed Vetted credential launch?

In build. Stage-1 Skill Passport is targeting Q4 2026. Stage-2 domain exams and Stage-3 paid calibration trials follow. Subscribe to the newsletter to be alerted for the first cohort — no waitlist gating.

What roles show up on HireFeed?

Three tracks in one feed. AI training (Tier 1): RLHF, LLM eval, agentic red-teaming, expert data, annotation. Freelance (Tier 2): engineering, ML, design, technical writing. US corporate (Tier 3): full-time roles at AI labs and enterprise data ops. All rate-disclosed.

Is my location eligible?

Depends on the platform. Most top platforms accept US, Canada, UK, Ireland, Australia, NZ. Some (Mercor, Surge, DataAnnotation) run global cohorts including India and select European markets. Eligibility maps by country are on the roadmap — until then, each listing carries its own eligibility line.

Frontier models need frontier humans.

Start with the feed. Sign up for the newsletter to be first in line for the HireFeed Vetted first cohort.

Preview mode  ·  Vetted, Skill Passport & sourcing product in build · Figures illustrative until production ships
Browse jobs