Reinforcement Learning Engineer (Fully Remote)

Unknown; United States; Remote; USD 120,000 - 180,000; Full time

Job summary

Unknown healthtech team builds clinical decision-support systems for public healthcare and seeks researchers to own reinforcement learning environments and evaluations ensuring reliable model behavior across multi-step clinical reasoning tasks. You will design reward functions, grading rubrics, and LLM judges, and run post-training experiments to identify failures and improve environments.

Proficiency in Python, PyTorch, and Transformers is required; experience with LLMs and reward modeling is a

Qualifications

• PhD in CS or ML, or at least 3 years of RL research or post-training experience.
• Experience building RL environments for post-training or agent evaluation.
• Proficient in Python, PyTorch, Transformers; familiar with LLMs and reward modeling.
• Strong evaluation and experimental design skills.

Responsibilities

• Own reinforcement learning environments and evaluations for reliable model behavior.
• Design reward functions, grading rubrics, and LLM judges; test for reward hacking.
• Run post-training and agent evaluation experiments and improve environments based on failures.

Skills

Python

PyTorch

Transformers

LLMs

Reward modeling

Reinforcement learning

Education

PhD in CS or ML

3+ years RL research experience

Tools

Gym

Transformers library

Job description

• We are building clinical decision-support systems for public healthcare.
• You will own reinforcement learning environments and evaluations for reliable model behavior.

What You'll Do:

• Build environments for complex clinical reasoning with multi-step tasks and tool use.
• Design reward functions, grading rubrics, and LLM judges, testing for reward hacking.
• Run post-training and agent evaluation experiments and improve environments based on failures.

What We're Looking For:

• PhD in CS or ML, or at least three years of RL research or post-training experience.
• Experience building RL environments for post-training or agent evaluation.
• Production-level Python, PyTorch, Transformers; familiarity with LLMs and reward modeling.

Unknown

We are a healthtech team building clinical decision-support systems for public healthcare. Our team emphasizes cross-disciplinary collaboration and rigorous evaluation practices.