Research — Poolside

The Model Factory. The system behind the models.

Traditional foundation model training is manual, linear, and slow. We built something different.

The Model Factory is Poolside's internal platform for training, scaling, and experimenting with foundation models. It handles automated evaluation during training, reinforcement learning from code execution, architecture ablations, synthetic data generation, and data mixing—all orchestrated across our GPU clusters.

Experiments that used to take weeks to schedule now run in under an hour. We describe a configuration, and the Factory handles the rest.

Inside the Model Factory

video: The Orchestration Multiplier

This video provides an overview of the "model factory" system at Poolside, highlighting its role as an end-to-end orchestration platform for building foundation models.

video: The 20% Decision

This video discusses the evolution of distributed training at Poolside, focusing on the shift in engineering strategy and the development of new infrastructure to support model building.

video: 1M Repositories

This video explains our use of reinforcement learning in our code execution environment, along with the ways we've scaled and automated processes in our Model Factory.

Explore further

The Model Factory Series A six-part blog series on how we train, evaluate, and iterate on foundation models. From GPU-to-GPU weight transfers and automated architecture ablations to reinforcement learning from code execution at scale.