Case Study: Actium Health achieves faster ML model training with Prefect

Actium Health cuts ML model training time by 99% with Prefect

Actium Health, a healthcare technology company, needed to scale its machine learning model training to rapidly develop a broad catalog of patient treatment recommendations. Facing significant time pressure, their initial use of the Dask library for parallel computing lacked the workflow orchestration required to automate and manage hundreds of experiments. This challenge led them to Prefect Core.

By implementing Prefect Core on a 100-node Dask cluster, Actium Health orchestrated parallelized pipelines, enabling data scientists to run hundreds of machine learning experiments quickly and reliably. Using Prefect Cloud provided enhanced visibility and monitoring. The solution accelerated research output, led to the production of successful models for oncology and other service lines, and cut model training time by 99%, all while maintaining compliance as sensitive data never passed through Prefect's infrastructure.

Actium Health