Staff Research Engineer - Frontier AI and Reinforcement Learning - International

What You'll Do8

• Probe what frontier models can and cannot do, shape research questions, and try out new training approaches.
• Assemble research datasets, experiments, prototypes, tooling, and evaluation frameworks from the ground up.
• Train, test, and evaluate models with Python and modern ML frameworks.
• Read results with scientific rigor, checking data quality, reproducibility, and the evidence behind each interpretation.
• Move from hypothesis to validated insight, then iterate on whatever holds up.
• Work with Research, Engineering, Product, and Operations to turn findings into practical systems.
• Share results through technical reports, publications, open-source work, and research forums.
• Review peers' work, mentor engineers and researchers, and carry technical leadership on the project.

Requirements9

• 7+ years of professional experience, with a substantial share in research engineering for machine learning or frontier AI systems.
• PhD or master's in AI, machine learning, computer science, or a neighboring technical field, or research experience strong enough to stand in for the degree.
• Solid ML foundations plus hands-on work designing experiments, training or evaluating models, or building AI systems.
• Research background in at least one area: synthetic or agentic data generation, reinforcement learning, post-training, model understanding, AI evaluation, benchmarks, or tool-using agents.
• Strong Python skills and comfort implementing, testing, and iterating with modern ML frameworks and tooling.
• Scientific judgment and reproducible methods, with close attention to data quality.
• Clear written and verbal communication, plus the ability to work independently and across functions.
• Experience mentoring engineers or researchers and providing technical leadership.
• Commitment to at least 40 hours per week, 8-hour days, and four hours of daily overlap with Pacific Time.

Who Should Apply

Researchers holding a PhD or master's with seven or more years in ML, including shipped experiments in reinforcement learning, post-training, or agentic data. You want a short, concentrated freelance engagement where findings feed real AI systems, and you can hold four hours of Pacific Time overlap each day.

Salary Insight