Software Engineer, RL Data
Core
Designing tasks, rewards, and environments to train frontier coding agents using reinforcement learning on real user data.
Role type
Software Engineer (RL Data)
Builds
Training datasets, task sets, reward systems, and evaluation environments for coding agents.
Domain
AI/ML, Reinforcement Learning, Software Engineering
Deliverable
production ML models
Required skills
Software engineering fundamentals, distributed systems, data engineering, task design, failure mode analysis, data quality improvement
Preferred skills
Reinforcement learning experience, research partnership
Technologies
N/A
Responsibilities
Designing task sets to teach specific agent capabilities, analyzing agent traces to identify failure modes, building systems to surface specific behaviors, converting one-off recipes into reusable tools, partnering with research on dataset effectiveness
Seniority
Mid-level, hands-on IC
