Model Engineer
Core
Research Scientist driving effective Reinforcement Learning (RL) and mid-training research for frontier coding agents and Super Intelligence stack.
Role type
Research Scientist (RL & Mid-training)
Builds
Frontier coding agents, RL environments, and evaluation tasks for the Super Intelligence stack
Domain
Artificial Intelligence, Reinforcement Learning, Large Language Models
Deliverable
production ML models
Required skills
Python, PyTorch, JAX, Rust, C++, RL algorithms, distributed training, model scaling, data quality, experiment design, reward design
Preferred skills
Ambiguous research ownership, data science, RL fundamentals, frontier-scale model training, original ML contributions
Technologies
Python, PyTorch, JAX, Rust, C++, Colossus supercomputers
Responsibilities
Design and run rigorous experiments to improve model capability; Improve data recipes through quality and deduplication; Define RL data and rewards including grader design; Advance training science from architecture to post-training; Build evals to expose model weaknesses; Partner with engineers to scale promising results
Seniority
Senior, hands-on IC
