Research, Coding Agents

Work on agentic coding research for frontier models by designing and running RL training jobs that improve long-horizon, multi-turn code writing, debugging, and reasoning. Own the coding post-training stack, including synthetic and human data generation, sandboxes, reward and grading design, and large-scale training runs. Build scalable data pipelines, design real-world evals, debug large RL runs, and collaborate with infra and post-training teams to ship improvements each model release.