Research Engineer, Benchmarks
About the Role
You will own the design and implementation of rigorous, domain-specific benchmarks used to evaluate frontier AI agents on realistic workflows. Sitting within a small, highly technical team of researchers and engineers, this role is central to delivering evaluations that AI labs and enterprise customers genuinely trust and rely on.
What You'll Do
Design, implement, and maintain high-quality internal benchmarks for evaluating frontier agents on domain-specific tasks.
Partner with subject-matter experts to define realistic workflows and translate them into well-scoped evaluation tasks.
Build and operate reliable infrastructure to run models and agents against benchmark tasks at scale.
Develop metrics and statistical analyses that measure benchmark difficulty, reliability, and failure modes.
Validate that benchmark performance correlates with real-world evaluations and customer needs.
Write clear technical documentation and benchmark reports for research and engineering audiences.
What We're Looking For
2 to 4 years of experience in research engineering or ML engineering, with a focus on AI benchmarks, evaluation infrastructure, or agent environments.
Strong proficiency in Python, Docker, and Linux for building research or production infrastructure.
Demonstrated experience designing and running benchmarks or evaluation environments for AI agents or large language models.
Experience building infrastructure to reliably run AI models or agents against evaluation tasks at scale.
Experience developing metrics or validation studies to assess benchmark difficulty, reliability, and real-world correlation.
Ability to collaborate with domain experts and translate complex workflows into evaluation criteria.
Strong attention to detail, with a habit of spotting subtle inconsistencies and edge cases.
Comfort working independently in fast-paced, early-stage startup environments with unstructured problem spaces.
Excellent written communication skills for technical documentation and cross-timezone collaboration.
Published papers or technical writing on AI benchmarking, model evaluation, or failure modes is a strong plus.
Experience with reinforcement learning training pipelines, data generation, or RL agent evaluation is a plus.
Compensation & Benefits
Salary range: $150,000 to $250,000 USD annually. Visa sponsorship is available.
Location
On-site in Singapore.