Epsilon Health
Research Engineer / Research Scientist, Post-Training & Reinforcement Learning
- Confirmed live in the last 24 hours
- No salary listed
- Mid level
- Full-time
- On-site · San Francisco, CA
- Added 1 month ago
About this role
Epsilon Health is seeking a talented Research Engineer or Research Scientist to advance the field of post-training and reinforcement learning for medical AI. This role focuses on developing innovative reinforcement learning methods and reward models to improve the accuracy and reliability of radiology reports generated by the company's foundation models. You'll work in a collaborative environment with direct access to real-world clinical data and feedback, bridging the gap between research and patient care.
What you'll do
- Advance reinforcement learning techniques for multimodal models.
- Develop and scale post-training methods using reinforcement learning.
- Design and train reward models incorporating clinical accuracy and radiologist feedback.
- Improve multimodal reasoning capabilities for medical image analysis.
- Explore agentic AI and tool use within imaging workflows.
- Own the research process from ideation to deployment.
What they're looking for
- Reinforcement Learning
- Language Model Post-Training
- Multimodal Machine Learning
- PyTorch or JAX
- Reward Modeling
- Preference Optimization
- Distributed Training
- Experimental Design
Benefits
- Freedom to pursue ambitious research ideas
- Resources for bringing ideas to life
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Epsilon Health
- Industry
- Technology & Software
Likely interview questions
- Describe a challenging reinforcement learning project you've worked on and how you overcame the difficulties.
- How would you design a reward model for training an AI to generate radiology reports?