CareerPlan

Research Engineer / Scientist - Reinforcement Learning

San FranciscoRemoteFull-time2026-10-16 → 2026-11-19
HybridVisa sponsorship offered

Core

Develop and test reinforcement learning (RL) algorithms, scaling research, and training frameworks for open superintelligence models focused on reasoning, coding, and long-horizon tasks.

Role type

Research Engineer / Scientist (Reinforcement Learning)

Builds

Open-source models and full-stack training infrastructure for frontier AI

Domain

Artificial Intelligence / Reinforcement Learning / Distributed Systems

Deliverable

production ML models

Required skills

RL algorithms, empirical research on model training, distributed ML systems, Python, PyTorch, experimental design, debugging model/system behavior

Preferred skills

distributed training frameworks, asynchronous rollout pipelines, diagnosing unstable RL runs, coding agents, verifiable rewards, long-horizon environments

Technologies

Python, PyTorch, prime-rl

Responsibilities

Develop RL methods for reasoning, coding, and long-horizon tasks; Investigate stability in asynchronous RL; Design experiments to understand RL performance scaling; Extend RL frameworks for new algorithms and workflows; Build diagnostics for unstable runs and performance regressions

Seniority

Mid-Senior, hands-on IC