Research Engineer / Scientist - Reinforcement Learning
$150k - $350k • Remote • San Francisco • FullTime
Posted 1d ago
Job Location
San Francisco
Tech Stack
Remote Work Policy
Fully remote
Employment Type
FullTime
Categories
AI Research Engineer
About the job
Prime Intellect is building the open superintelligence stack, a full-stack system for post-training at frontier scale, from SFT and RL to tool use and agent workflows. We are developing open-source models trained end-to-end for long-horizon tasks like autonomous research, and the platform our own research team uses to build them. We are seeking individuals passionate about building at the intersection of frontier research, real infrastructure, and go-to-market for a nascent category. This role will focus on developing the algorithms and systems used to train our open models with reinforcement learning, with opportunities to specialize in RL algorithms, scaling research, or training frameworks.
Responsibilities
- Develop and test RL methods for reasoning, coding, and long-horizon tasks, including changes to training objectives and policy optimization.
- Investigate stability in asynchronous RL, including policy staleness, trainer/sampler mismatch, and numerical differences.
- Collaborate with data and evaluation researchers to understand the impact of task selection, rewards, and training data on model learning.
- Design experiments to analyze how RL performance and learning dynamics are affected by model size, training compute, and task horizon.
- Use controlled ablations and evaluations to differentiate learning gains from reward exploitation, evaluation artifacts, or training bugs.
- Translate experimental findings into training recipes and validate their effectiveness across larger models and longer runs.
- Extend prime-rl to support new algorithms, models, and asynchronous training workflows.
- Enhance the performance and numerical correctness of research runs across trainers, rollout workers, and environments.
- Build diagnostics to facilitate reproduction of unstable runs and performance regressions, distinguishing algorithm issues from implementation or runtime problems.
Requirements
- Demonstrated depth in at least one of RL algorithms, empirical research on model training, or distributed ML systems.
- Strong programming skills in Python and experience with PyTorch or a comparable ML framework.
- Ability to design controlled experiments, interpret results, and debug unexpected model or system behavior.
- Experience taking a research idea or systems change through implementation, testing, and evaluation.
- Experience with distributed training frameworks or asynchronous rollout pipelines (nice to have).
- Experience diagnosing unstable RL runs, including policy-staleness or numerical issues (nice to have).
- Experience with coding agents, verifiable rewards, or long-horizon environments (nice to have).
Benefits
- Cash Compensation Range of $150-350k
- Equity incentives
- Flexible work arrangements
- Option to work remotely or in-person
- Visa sponsorship
- Relocation assistance
- Quarterly team off-sites
- Hackathons
- Conferences and learning opportunities