All open roles

AI

AI evaluation and measurement

Active opening

Careers

Research Evaluation Engineer

Ensure that a better benchmark score represents a meaningful improvement.


The role

Responsibilities
Build evaluators and comparable baselines. Detect invalid outputs and evaluator exploitation. Maintain experiment provenance and reproducibility. Work with domain researchers on correctness criteria and with AI researchers on ablations and cost measurements.
Relevant experience
Evaluation systems, testing, benchmarking, scientific software, or adversarial analysis of automated workflows.
Evidence to share
An evaluation design that changed the interpretation of a claimed result.

How to apply

Email Ivan Miskovic with the role title in the subject line. Include a CV or profile and links to relevant work: a paper, an implementation, a technical note, or a project you can explain in detail. Do not send private keys, sensitive identity documents, salary history, or private research materials.

Apply by email

miskovic@potomac.net

Location, working arrangement, and compensation are discussed directly with candidates.