Case Study: Medely achieves real-time healthcare staffing and faster ML velocity with Chalk

Medely staffs critical healthcare in real-time with Chalk

Medely, a healthcare staffing marketplace, faced significant challenges scaling its machine learning initiatives for real-time pricing and professional matching. Their batch pipeline limited feature freshness to 24 hours and required heavy manual infrastructure work for any changes, slowing experiment cycles to approximately two months. They turned to the vendor Chalk and its self-serve feature platform to overcome these obstacles.

By implementing Chalk, Medely replaced its fragile batch system with a real-time feature computation infrastructure. This self-serve solution eliminated manual engineering overhead, allowing feature changes to be made in minutes instead of days. The results were substantial; their first model deployed with Chalk is projected to generate $800,000 in annual net revenue. The platform also accelerated the team's experiment velocity, moving them from two-month cycles toward two-week sprints.

Medely

Eric Simon

Staff Machine Learning Engineer