Reinforcement Learning (RL) Researcher

Main mission

Develops reward-based learning methods that drive agents and robots.

5 key responsibilities

• Design reward learning algorithms and their environments.
• Define robust and non-circumventable reward functions.
• Train and evaluate agents on complex sequential tasks.
• Study exploration, generalization, and policy stability.
• Transfer RL methods to robotics, games, and LLMs.

Key skills

RL (PPO, Q-learning, RLHF), simulation, optimization mathematics, PyTorch.

What's expected

Agents that learn useful behaviors without exploiting weaknesses in their reward.

Career paths

Post-training Engineer, Robotics Researcher, RL Research Lead.