Research Scientist/Research Engineer, Midtraining

Sourced from radical.getro.com

Periodic Labs is an AI and physical sciences company training frontier models to accelerate breakthroughs in materials, energy, and beyond. As a Midtraining Research Engineer, you will take base models and improve their scientific reasoning by curating and generating data, building evaluations, and running large-scale training experiments. You will work closely with RL researchers, physicists, and chemists, and lay the groundwork for pre-training efforts.

• Identify, process, and curate novel sources of scientific data for large-scale model training.
• Generate high-quality synthetic data to fill gaps in scientific knowledge and reasoning.
• Build evaluations that correlate with downstream scientific task performance.
• Develop and apply techniques such as self-distillation and on-policy distillation to improve model capability.
• Design and run large-scale training experiments, partnering with supercompute engineers to scale efficiently across thousands of GPUs.
• Build tools for yourself and the team to investigate how data choices shape model intelligence.

• Experience training LLMs on curated mixes of trillions of tokens.
• Experience on a dedicated evals team supporting a large production training run.