About the Role Join a small, technical team working on AI evaluation and reinforcement learning infrastructure. You will design and build rigorous benchmarks for assessing AI agents on realistic, domain-specific workflows, helping research and engineering teams understand how well agents perform in practice. What You'll Do • Design, implement, and maintain benchmarks for evaluating AI agents on domain-specific tasks. • Work with subject-matter experts to translate real workflows into practical tasks and evaluation criteria. • Build reliable infrastructure to run models and agents against benchmark tasks at scale. • Develop metrics and analyses to assess benchmark difficulty, reliability, and failure modes. • Validate benchmark results against real-world evaluations and user needs. • Write clear technical documentation and reports that explain methods and results. What We're Looking For • Two to four years of experience in software engineering, machine learning engineering, or research, including at least two years building AI benchmarks, evaluation infrastructure, or agent environments. • Hands-on experience designing and running evaluations or benchmark environments for AI agents or large language models. • Strong proficiency with Python, Docker, and Linux, including building and operating reliable infrastructure. • Experience developing metrics, statistical analyses, or validation studies for evaluation systems. • Experience collaborating with subject-matter experts and understanding workflows across technical or business domains. • Strong technical writing, careful attention to detail, and the ability to work independently in an early-stage, fast-moving environment. • Research publications or technical writing on benchmarking, model evaluation, or failure modes are beneficial. Experience with reinforcement learning pipelines or public benchmark projects is also useful. Compensation & Benefits Base salary: $100,000 to $170,000 USD annually. Visa sponsorship is available. Location On-site in Singapore, Singapore.