Expert Talent Pool, Evals and Benchmarks

CleraRemoteFull-timeSenior, 5–8 yearsListed 15 minutes ago

Apply now

About this role

About the Role

Join a network of senior evaluation experts supporting AI research projects. When projects match your experience, you may lead benchmark design, validation, and grading operations to help assess model performance rigorously.

What You'll Do

- Design benchmarks from the ground up, defining scope, question types, difficulty levels, and scoring methods.
- Write and review rubrics to promote consistent grading and strong agreement across evaluators.
- Develop gold-standard answers and test sets that can withstand expert review.
- Lead grading rounds by calibrating graders, measuring agreement, and resolving disputes.
- Analyze model failures, identify root causes, and recommend improvements.
- Design and conduct rigorous model comparison studies.
- Validate benchmarks for contamination, ambiguous answers, and saturation.
- Design user studies on how people rate and interact with model outputs.

What We're Looking For

- At least 5 years of experience in evaluation, testing, or assessment, or 5 years of domain expertise combined with hands-on evaluation work.
- Experience designing a benchmark, test plan, exam, or assessment used by others.
- Experience writing or reviewing rubrics and gold-standard answers, and reviewing or leading graders or question writers.
- Proficiency with Python and evaluation tools such as Inspect, lm-eval-harness, HELM, or promptfoo.
- Experience analyzing model or test failures and recommending fixes.
- Background in machine learning evaluation, QA, test engineering, psychometrics, assessment design, UX research, data science, or subject matter expertise.
- Clear written English communication. Experience validating benchmarks, measuring grader agreement, running model comparisons, or conducting user studies is relevant.

Compensation & Benefits

Compensation is $75 to $175 per hour, depending on project, field, and experience level. Payment is based on hours worked.

Location

Fully remote. Projects may last from one week to several months.