HiringSenior Coding Experts earn up to $150/hr→
HIRING Coding experts up to $150/hr

The experts
behind frontier AI.

HirePower AI supplies expert human data for code and agentic AI. Practising engineers and domain specialists write, review and stress-test the work that trains and evaluates the world's most capable models.

50K+
Hired & placed
$150
Top hourly rate
100%
Remote
Weekly
Payouts
A global expert network supporting AI work across the industry, including teams at
LinkedIn Microsoft MAI & more
01
For experts

Apply as an expert.

Remote, flexible work evaluating and improving frontier AI. Coding Experts earn up to $150/hr, Generalist Experts $100+/hr, paid weekly.

Open roles

Get paid to make AI smarter.

Coding Expert

Software engineering · Agentic coding
Hiring
Hourly rate up to$100 / hour

0–2 years of professional experience. Strong fundamentals in at least one language and a sharp eye for bugs.

PythonTypeScriptJavaGoRustC++SQL
Location
Remote
Commitment
Flexible
Type
Contractor
Payouts
Weekly

Generalist Expert

Reasoning · Writing · Domain knowledge
Hiring
Hourly rate$100+ / hour

Evaluate and improve AI responses across reasoning, writing, research and fact-checking. A degree or deep expertise in your field and excellent written English.

ReasoningWritingFact-checkingMathScienceFinanceLaw
Location
Remote
Commitment
Flexible
Type
Contractor
Payouts
Weekly
Work when you wantNo minimum hours. Pick up tasks on your schedule.
Paid weeklyTransparent hourly rates, reliable weekly payouts.
Fully remoteWork from anywhere with a laptop and a connection.
Grow with qualityTop reviewers unlock senior and adjudicator work.
How it works

From application to first payout.

No agencies, no middlemen. Apply once, then do everything else on the HirePower AI platform.

02
For AI teams

Experts and data for frontier models.

Two ways to work with us: staff your projects and pipelines with vetted domain experts, or hand us a data annotation project end to end.

Expert PoolsData Annotation
Best whenYou have tooling and a team running the project, and need the right expertsYou want a finished, quality-checked dataset delivered to spec
Who runs the workYour team, with our pool leadHirePower AI, end to end
You getCalibrated experts working inside your pipelineAnnotated data, QA reports and guideline changelog
We handleSourcing, vetting, calibration, scheduling, payments and complianceAll of that, plus guidelines, tooling, QA, adjudication and delivery
0.91κ
Agreement on SWE task authoring, rolling 30 days
6
Quality gates on every item before delivery
50K+
People hired & placed across engineering and expert domains
100%
Human work, with a contractual no-model-assistance rule
02.A
Expert Pools

Domain experts, on tap for your pipelines.

Whether you need twenty Rust engineers for an RL environment sprint or a standing bench of clinicians for safety evals, we build the pool, calibrate it to your guidelines and keep it performing.

Domains

Who's in the pool.

Specialties
Typical work
    How we vet

    Vetting

    Only a fraction make it to your pool.

    Application & profile review
    Domain assessment
    Practical or live test
    Training courses
    Calibration on your gold set
    Your pool

    Every expert clears each stage before they touch your data. Calibration runs on your own guidelines and gold set.

    How a pool comes together

    Brief to productive, fast.

    01Brief

    Skills, seniority, languages, time zones, weekly hours and the tools experts will work in.

    02Match

    We shortlist from our network and recruit for anything niche.

    03Calibrate

    Onboarding on your guidelines and gold set. Only experts who hit the bar go live.

    04Deploy & scale

    Experts work in your pipeline. We track performance, replace and scale week to week.

    What's included

    You get the experts. We carry the overhead.

    Sourcing & vettingMulti-stage screening by senior experts in each domain.
    Onboarding & calibrationTraining on your guidelines, measured against your gold set.
    Pool leadOne point of contact for scheduling, questions and escalations.
    Performance trackingQuality and throughput per expert, with fast replacement.
    Contracts & paymentsGlobal contractor agreements and weekly payouts, handled.
    Security & NDAsIndividual NDAs, least-privilege access and audit logging.
    Pool models

    Shape the pool around your roadmap.

    On-demand pool

    Flexible capacity

    A vetted bench that ramps up and down with bursty pipelines and launch deadlines.

    Dedicated pod

    A team that knows your models

    Named experts with a pool lead and fixed weekly capacity, building context over time.

    Embedded specialists

    Senior experts beside your researchers

    Domain leads who help write guidelines, design tasks and review edge cases with your team.

    02.B
    Data Annotation

    Managed annotation, from guidelines to gold.

    Hand us the outcome you need. We write or refine the guidelines, staff the right experts, run every quality gate and deliver data your researchers can trust.

    Modalities

    Annotation across every data type.

    { }CodeTask authoring, review, preference ranking, bug labelling
    ↺Agent trajectoriesStep-level verdicts, error localisation, reward-hacking flags
    AaTextSFT writing, RLHF ranking, fact-checking, classification
    ▣ImageCaptioning, bounding boxes, segmentation, VQA
    ∿Audio & speechTranscription, diarisation, speech quality and emotion
    ▶VideoEvent tagging, temporal segmentation, video QA
    ◫DocumentsOCR correction, layout, extraction, table parsing
    ✦MultimodalCross-modal grounding, preference and safety evaluation
    Workstreams

    Where we go deepest.

    01

    SWE task authoring

    Graded engineering tasks with test harnesses and reference solutions, built from real-world repos.

    02

    Agentic trajectory evaluation

    Step-level review of multi-turn, tool-using agent rollouts — catching false success claims and silent failures.

    03

    Code preference & RLHF

    Ranked response pairs with written rationales. Disagreements go to senior adjudication, not a coin flip.

    04

    Evals & red-teaming

    Held-out benchmarks, contamination screening and adversarial probing to find where models really break.

    Sample deliverables

    See the work, not a slide about it.

    ILLUSTRATIVE SAMPLE
    
          
    What you receive
      Quality gates

      Samples are illustrative. Client work is delivered under NDA and never shown publicly.

      Quality system

      Six quality gates on every annotation project.

      01Source & vet

      Referral-led sourcing, live coding or domain screens, and assessments scored against gold data.

      02Calibrate

      Joint guideline sessions with your researchers and a shared gold set before production starts.

      03Annotate blind

      Overlapping, blind assignments so agreement is measured, not assumed.

      04Auto-check

      Tests executed, schemas validated, leakage and duplicate screens run on every item.

      05Adjudicate

      Disagreements escalate to a senior reviewer, who writes a ruling and feeds it back into the guidelines.

      06Deliver & report

      Data lands in your environment with per-workstream agreement and error reports.

      Why HirePower AI

      Built for the annotation work generic vendors can't do.

      HirePower AICrowd vendorsIn-house only
      Who does the workPractising engineers & vetted expertsGeneralist annotatorsYour own engineers
      Code executed & verified✓Rarely✓
      Agreement tracked per workstream✓VariesRarely
      Senior adjudication✓—Ad hoc
      Works inside your tooling✓Their platform✓
      Scales without hiring✓✓—
      No model assistance, by contract✓Varies—
      Deliverables

      What lands in your bucket.

      Annotated dataJSONL, Parquet or your own schema, versioned per batch
      Agreement reportInter-rater agreement per workstream and label
      Error taxonomyWhere and why items failed QA, with examples
      Guideline changelogEvery edge-case ruling, written down
      Weekly statusThroughput, quality trends and open questions
      The frontier of AI data

      Measured, not assumed.

      Model progress now hinges on harder, verifiable, expert-made data. Here's where the field is heading — and how we hold ourselves to it.

      Quality signal · Preference agreement

      Calibration compounds.

      Code-preference agreement rose from 0.61 on day 1 to 0.72 by day 30 through calibration rounds, blind re-annotation and adjudication.

      Day 1 · 0.61Day 15Day 30 · 0.72
      Rolling 30-day agreement

      By workstream

      SWE task authoring0.91
      Trajectory evaluation0.84
      Code preference0.72
      Benchmarks our work maps to

      Where models are judged

      SWE-bench Verified REPO-LEVEL SWE
      Terminal-Bench AGENTIC CLI
      τ-bench TOOL USE
      LiveCodeBench CONTAMINATION-FREE
      GPQA Diamond EXPERT Q&A
      Trend · Agents

      From answers to actions

      Labs are shifting from single-turn chat data to long-horizon agent trajectories, where one bad tool call can sink a 40-step run. Reviewing them takes engineers, not checklists.

      Trend · Verifiable rewards

      RL environments

      Reinforcement learning needs tasks with checkable outcomes: unit tests, graders and sandboxed environments. Our experts design tasks that are hard to game.

      Network
      50K+

      People hired and placed across engineering and expert domains, trusted by teams at leading AI labs.

      Join them →
      PythonTypeScriptRustGoJavaC++KotlinSwiftSQLKubernetesTerraformReactDjangoPyTorchDistributed systemsSecurityMathematicsPhysicsFinanceLawMedicineTechnical writing
      Security & trust

      Your data stays yours.

      Work happens inside your environment — your tooling or an internal platform — under individual NDAs and audited access.

      Talk to us →
      Individual NDAsEvery contributor, every engagement
      In place
      Access control & audit logsLeast-privilege access, fully logged
      In place
      No-model-assistance ruleContractual: human work only
      In place
      Ways to engage

      Start small. Scale what works.

      Pilot

      Prove it on your data

      • Expert pool or annotation, one scope
      • Shared gold set
      • Agreement and quality report
      • Clear go / no-go decision
      Request a pilot
      Managed annotation

      A finished dataset, delivered

      • Guidelines, workforce, QA and delivery
      • Six quality gates on every item
      • Agreement and error reports
      • Dedicated project lead
      Request a pilot
      Start a pilot

      Tell us what you're building.

      Need experts for your pipeline, or a dataset delivered end to end? We'll come back with a scoped pilot: who we'd staff, the quality bar and how we'll report against it.

      03
      OTS Datasets

      Buy or sell datasets.

      We work with data partners and data holders to license rights-cleared, off-the-shelf datasets to AI teams.

      Data acquisition & licensing

      Off-the-shelf datasets, rights-cleared.

      Alongside custom data, HirePower AI works with data partners and data holders to source, license and deliver off-the-shelf (OTS) datasets to AI teams. Buyers get ready-to-use data with clear provenance. Data holders get a trusted route to market for their data.

      SourcePartnerships with data holders, publishers and platforms
      Clear rightsLicensing, consent and provenance documented
      PrepareDe-identification, cleaning, deduplication and QA
      LicenseFlexible terms and secure delivery to buyers
      Dataset categories
      Code & repositoriesAgent & tool-use trajectoriesInstruction / SFTPreference & RLHFConversational & dialogueReasoning & mathSTEM & scientificMultilingual textWeb & news corporaBooks & long-formDocuments & OCRSpeech & audioMusicImage & captioningVideoHealthcare (de-identified)LegalFinancialE-commerce & productGeospatial & satelliteRobotics & sensorAutonomous drivingUI & screen interactionSynthetic dataEvaluation sets
      Reach out

      Buying or selling OTS data?

      Or email [email protected]

      FAQ

      Questions, answered.

      Still have questions? Email [email protected]

      Get started

      Wherever you fit, start here.

      Join our expert network, bring us your hardest data problem, or buy and sell datasets through a partner you can trust.

      Trust & Safety

      Stay safe from recruitment scams.

      Your safety online matters to us. Fraudsters sometimes pretend to be well-known AI training companies, and they may use our name too. These guidelines will help you recognise genuine contact from HirePower AI.

      We will never ask you to pay us.

      We pay you; you never pay us. There are no fees for applying, assessments, training courses, equipment or "activation". We only ask for payment details inside our platform while you're logged in, and only to pay you. If you're not logged in, don't give financial details to anyone who says they're from HirePower AI.

      Only talk to our official recruitment team.

      Only reply to messages from our verified recruitment team. If anyone else contacts you directly outside our official channels, don't reply. Our emails come from the hirepowerai.tech domain only, so check the sender's full address before you respond.

      Apply only through our official sign-up flow.

      The only way to apply is the "Apply now" button on this website. We don't accept applications any other way. If someone invites you to apply outside our standard onboarding process, ignore it.

      Signs that a message isn't from us

      • It asks you to pay anything, or to buy gift cards, crypto or equipment.
      • It comes from a personal email address or a domain other than hirepowerai.tech.
      • It sends you a link to apply that doesn't go through this website.
      • It pushes you to decide quickly, or offers pay far above the rates listed on our Opportunities page.
      • It asks for passwords, one-time codes or bank details outside our logged-in platform.
      Think someone is pretending to be us?Don't reply and don't send money. Forward the message to us so we can look into it.
      [email protected]
      Legal

      Privacy Policy

      Last updated: November 2026

      HirePower AI ("HirePower AI", "we", "us") connects skilled experts with AI training and evaluation work. To do that, we handle personal data. This policy explains what we collect, why we collect it, who we share it with and the choices you have. It covers our website, our platform and any other way you deal with us (together, the "Services").

      1. Data we collect

      What we collect depends on how you use the Services and on whether you work with us as an independent contractor.

      • Identity and contact details: your name, email address, phone number, country of residence and account login details. If you work with us as a contractor, we may also collect details needed to verify your identity and meet tax and legal requirements.
      • Professional and education details: your CV, work history, skills, programming languages, qualifications, area of study and any professional profile links you choose to share.
      • Assessment and work data: your answers to assessments and training courses, scores, the tasks you complete, quality and agreement metrics, and time spent on tasks.
      • Communications: messages, feedback and support requests you send us.
      • Device and usage data: IP address, browser and device type, operating system, pages viewed, links clicked and approximate location inferred from your IP address.
      • Payment data: the details needed to pay you. You enter these inside our logged-in platform, and our payment providers may process them directly.
      • Optional information: any extra information you choose to give us, for example in a survey. Where the law treats this as sensitive data, we only process it with your explicit consent.

      Clients who contact us about our services give us business contact details, such as name, work email, company and role.

      2. Where we get your data

      • From you: when you apply, create an account, take assessments, complete tasks or contact us.
      • Automatically: through logs, cookies and similar technology when you use our website and platform.
      • From third parties: for example identity verification, analytics, advertising and recruitment partners, and people who refer you to us, where the law allows.

      3. How we use your data

      • To review your application, run assessments and training, and match you with suitable projects.
      • To provide and run your account, assign and review work, and pay you.
      • To check quality, prevent fraud and enforce our rules. That includes our ban on using AI models to do the work.
      • To contact you about your account, available opportunities, security alerts and support.
      • To improve our Services through analytics and internal research.
      • To deliver client projects. The work you produce may be used to train, evaluate and improve AI models.
      • To meet legal, tax and regulatory obligations, and to establish or defend legal claims.

      We won't use your data for purposes that don't fit with these unless we tell you first and, where required, get your consent.

      4. Legal bases for processing

      Where the law requires a legal basis, we rely on one or more of these:

      • Contract: to carry out our agreement with you.
      • Legitimate interests: for example running and improving our Services, keeping them secure and preventing fraud, as long as your rights don't override those interests.
      • Consent: for example optional marketing or sensitive information. You can withdraw consent at any time.
      • Legal obligation: where we must process data to comply with the law.

      5. Who we share your data with

      We do not sell your personal data. We share it only as follows:

      • Service providers who help us run the business, such as hosting, analytics, identity verification, communications and payment providers. They may use your data only to provide services to us.
      • Clients whose projects you work on. They receive the work you produce and, where needed, limited profile information such as your skills, qualifications or an anonymised contributor ID.
      • Recruitment and advertising partners, in limited form, to help us reach people who may want to join us.
      • Authorities, when the law requires it, or to protect the rights, property or safety of HirePower AI, our users or others.
      • A buyer or successor, if we're involved in a merger, acquisition or sale of assets.

      6. International transfers

      We work with contributors, clients and service providers in several countries, so your data may be handled outside the country where you live. When we move data across borders, we use the safeguards the law requires, such as standard contractual clauses.

      7. How long we keep data

      We keep personal data only as long as we need it for the purposes in this policy. For contractors, that means while you work with us and afterwards for as long as legal, tax, accounting or dispute-resolution needs require. After that, we delete or anonymise it. Work you deliver to a client as part of a project may stay with that client under their agreement with us.

      8. Security

      We use technical and organisational safeguards to protect personal data. These include encryption in transit and at rest, access controls with audit logging, individual NDAs for contributors and staff training. No system is perfectly secure. If a breach affects your data, we'll notify you and the relevant authorities where the law requires.

      9. Your rights

      Depending on where you live, for example under the EU and UK GDPR, India's Digital Personal Data Protection Act or US state privacy laws such as California's, you may have the right to:

      • Access the personal data we hold about you and get a copy of it
      • Correct data that is inaccurate or incomplete
      • Delete your data, subject to legal exceptions
      • Restrict or object to certain processing, including direct marketing and profiling
      • Receive your data in a portable format
      • Withdraw consent at any time
      • Opt out of targeted advertising or decisions made solely by automated means
      • Appoint someone to act for you, and complain to your local data protection authority

      To use these rights, email [email protected]. We may need to verify your identity first. We'll respond within the time the law requires, and we won't treat you differently for using your rights.

      10. Cookies and tracking

      We use cookies and similar technology to keep you signed in, keep the Services secure, understand how they're used and measure our recruitment campaigns. You can control cookies in your browser settings. If you block essential cookies, parts of the Services may stop working.

      11. Other information

      • Age: you must be 18 or over to use the Services. We don't knowingly collect data from anyone under 18.
      • Third-party sites: this policy doesn't cover websites or services we link to. Please read their own policies.
      • Changes: we may update this policy. If we make major changes, we'll tell you, for example by email or on our website.

      12. Contact us

      For questions, concerns or requests about this policy or your personal data, contact us at [email protected].

      Apply now
      Free to apply · about 5 minutes

      Apply to work with us.

      Tell us a little about you. We email your HirePower AI platform invite straight away, where you complete the assessments and training.

      There is never a fee to apply. Official emails only come from hirepowerai.tech.