Skip to content
Contributor Network

Analysts on a named pod, not IDs in a queue

India produces vast engineering and STEM talent that currently reaches AI labs as unmanaged freelancers with high variance. We organise that supply into vetted, trained, calibrated teams and pay it properly.

Pass rate
~10%
Reviewer ratio
1 : 6
Gold seeding
3–5%
Calibration
Weekly
Contributor funnelIndicative shape, not a quota

Selection is the quality strategy. Paid training and a real ladder from contributor to reviewer to lead is what keeps churn, and therefore recalibration cost, low.

Selection

Roughly one in ten applicants reaches the bench.

The pass rate is itself a client-facing selling point. Every stage after screening is paid, because asking people to work for free to prove they can work selects for the wrong things.

  1. Source100%

    Campus networks, referrals, engineering and Kaggle communities, IIT/NIT/IIIT alumni groups.

  2. Screen~30%

    Structured application plus an async, domain-matched skills test.

  3. Assess~40%

    A paid trial batch on gold tasks and a live rubric interview.

  4. Onboard~90%

    KYC, NDA and a paid calibration week before any client work.

  5. Bench~10% e2e

    Skills-tagged, score-tracked and deployable. Roughly one in ten applicants.

What you get

Contributors are professionals here.

Low churn is a quality strategy, not an HR nicety. A pod that stays calibrated is worth more than a pod that is cheap.

  • Paid training

    Trial batches and the calibration week are paid. You are never asked to work for free to prove you can work.

  • Transparent pay

    Top-quartile Indian technical freelance rates, published for the role, paid on time every cycle.

  • A real ladder

    Contributor → reviewer → senior reviewer → project lead. Most project leads are promoted from the contributor base.

  • Named, not anonymous

    You are a Newbieget Labs analyst on a named pod with a lead who knows your work, not an ID in a queue.

  • Weekly feedback

    A personal quality dashboard with your acceptance rate, gold-task accuracy and review outcomes. No black-box scoring.

  • Real work

    Frontier evaluation, agent benchmarking, RLHF for teams building the models, not bounding boxes.

Promotion ladder

A path no gig platform offers.

The ladder is the retention engine. Most project leads are promoted from the contributor base, because they already know the rubrics.

  1. 01ContributorReviewer

    Sustained personal acceptance rate and gold-task accuracy above the floor.

  2. 02ReviewerSenior Reviewer

    Review accuracy against audit; demonstrated coaching of contributors.

  3. 03Senior ReviewerProject Lead

    Rubric authorship and ownership of a QA gate across a full program.

  4. 04Project LeadHead of Delivery

    Program-level SLA ownership across multiple concurrent clients.

Expert bench domains
  • Software Engineering

    Python, JS/TS, Go, Rust, systems

  • Machine Learning

    Training pipelines, evaluation methodology

  • Medicine

    Practising clinicians for clinical-safety review

  • Law

    Qualified practitioners for legal-reasoning tasks

  • Finance

    CAs and analysts for quantitative review

  • Mathematics

    Proof verification and formal reasoning

  • Sciences

    Physics, chemistry, biology at graduate level

  • Linguistics

    Multilingual evaluation and annotation

Applications

Technical contributors, reviewers and expert-bench specialists. Screening is a structured application plus a domain-matched skills test.