STEM PhD evaluators
Mathematics, physics, chemistry, biology and microbiology. PhD holders who write and grade graduate-level reasoning tasks.
Talent bench
Vetted experts for AI training and evaluation.
For AI labs and data vendors. We recruit, interview and verify domain experts and native-speaker linguists, then send you a shortlist with the evidence behind each person.
Weekly-paid contractors · Proctored interviews · Verified credentials · 16 roles hiring now
Role families
Who is on the bench
The role families AI teams ask us for most. Each is vetted against the credentials that matter for that field.
Finance experts
CFA and CPA holders, investment banking, private equity, credit, FP&A, accounting and insurance practitioners.
Lawyers
Bar-admitted attorneys, including US-licensed, for legal reasoning, drafting and rubric design.
Physicians and specialists
Licensed physicians and board-certified specialists, including radiology and dermatology, for clinical reasoning evaluation.
Software engineers
Senior engineers in Python and C++ for agentic, long-horizon coding tasks and RL environment authoring.
AI safety and red-teaming
CBRN background, cyber threat intelligence, OSINT, jailbreak and prompt-injection testing, child-safety research.
Arabic dialect linguists
Gulf and Najdi, Levantine, Egyptian, Maghrebi and Iraqi. Native speakers who write their dialect accurately.
Low-resource languages
QA, transcription and text validation in South and Southeast Asian, African and European languages, by native speakers.
12 roles hiring now →Engineering domain experts
Energy and grid operations, mechanical engineering and data engineering practitioners.
Evaluation QA leads
Leads for agentic and RAG evaluation: rubric design, reviewer calibration and inter-annotator agreement.
How we vet
Evidence for every expert, not just a CV.
-
AI interview in English, Arabic or Urdu
A structured, role-specific interview in the language set for the role, scored against a rubric.
-
Full proctoring
Camera, screen and integrity checks throughout the interview, with a proctoring score on record.
-
Identity verification
Government ID and liveness check through Didit, with duplicate-account detection.
-
Registry credential checks
Licences, bar admission, board certification and degrees checked with the issuing body where a public register exists.
-
Expert Score tiers
Interview, domain test and trial results combine into an Expert Score and a tier: Vetted, Expert or Outlier.
-
Human second-chance review
Borderline candidates get a peer interview with a senior practitioner in the same field before any decision.
Coverage
Experts where you need them
Remote and global by default. Region-locked roles, such as US bar admission or North American generalists, are sourced only in that region.
- North America
- Europe
- Gulf
- South Asia
- Africa
Engagement
Simple to start, simple to scale.
Weekly-paid contractors
Experts are contractors paid in USD every week. You get one invoice.
Fast shortlists
Tell us the role, domain and languages. We send a shortlist with an evidence pack for each expert.
Replacement and trial terms
Replacement if an expert is not a fit, and trial work measured against your quality bar.
Talent bench
Tell us who you need. We send the shortlist.
Share the field, languages, volume and timeline. We post and vet the role at no cost to you.