Robotics and RL Researcher (Research Coding Tasks)

What You'll Do7

• Turn a published paper with public code into a research coding task an AI agent can attempt end to end.
• Remove the code that carries the paper's central contribution, then package the remaining project as the starting point.
• Write the task statement that tells an agent what to rebuild and which results it must match.
• Define held-out evaluation workloads and per-metric pass thresholds for the task.
• Supply a hidden reference solution plus a second working solution that clears the bar you set.
• Review agent attempts against your thresholds and record where each run holds up or falls short.
• Confirm the task stays solvable while still demanding real work from a capable agent.

Requirements6

• Published research in robotics or reinforcement learning at venues such as CoRL, RSS, ICRA, IROS or NeurIPS, or comparable standing.
• Released code for at least one paper, with a GitHub repository or artifact link you can share.
• PhD in computer science or a related field, finished or in progress.
• First-author papers that report measured gains over a baseline stand out.
• Comfort with Docker and benchmark setups that another person can reproduce.
• Depth in robot learning, sim-to-real transfer, control or reinforcement learning.

Who Should Apply

A PhD researcher or research engineer who has shipped code alongside a robotics or RL paper and can judge what makes a result reproducible. You will design evaluation criteria that hold up when someone else runs them.

Salary Insight