Full-time position
Research Engineer / Scientist - Reinforcement Learning
Prime Intellect
- San Francisco, United States
- Reinforcement Learning, Machine Learning and
About
Develop the algorithms and systems behind reinforcement learning for open models. Research tracks include long-horizon policy optimization, asynchronous training stability, and how learning changes with compute, model size and task horizon. The work includes controlled experiments, diagnostics and improvements to the training framework.
Requirements
Demonstrated expertise in reinforcement learning, empirical model-training research or distributed ML systems. Strong Python and a framework such as PyTorch. Ability to design controlled experiments and carry a change through implementation and evaluation.
Topics
- policy optimization, search World Wide
- Reinforcement Learning, search World Wide
- open models, search World Wide
- long-horizon policy optimization, search World Wide
- asynchronous training, search World Wide
- training stability, search World Wide
- compute, search World Wide
- model size, search World Wide
- task horizon, search World Wide
- controlled experiments, search World Wide
- training framework, search World Wide
Source: Prime Intellect