Periodic Labs

Drive mid-training research for frontier models that develop scientific reasoning, from curating and generating training data to building evaluations and running large-scale experiments. You’ll work across data, model training, and scientific domains to advance AI for discovery.

Build the high-performance systems layer powering Periodic Labs’ frontier model training, inference, and RL feedback loop for scientific discovery. You’ll work across GPU scheduling, kernels, RDMA, distributed infrastructure, and research co-design.

Build the full-stack orchestration platform that schedules instruments, manages synthesis workflows, and tracks scientific data across an automated materials lab. You’ll work in Python, React, and cloud infrastructure alongside scientists to make experiments reliable and reproducible at scale.

Build end-to-end agentic products that help materials scientists plan experiments, run analyses, and make decisions alongside AI agents. As an early product hire, you’ll define the interface between scientists, frontier models, scientific data, and lab systems.