Union.ai explains how it kept Python as the primary language for users while leveraging Rust to optimize performance-critical infrastructure. The piece details Flyte 2.0’s architecture, which uses Rust across the SDK, container runtime, data plane, and controller to handle workloads of tens of thousands of short-lived tasks and multi-gigabyte artifacts shuttled across regions, achieving significant speed improvements without sacrificing the Python developer experience.
How We Kept Python and Got Our Speed from Rust
calendar_today
June 15, 2026
person
Yee Hing Tong, Ketan Umare, Nary Yeh, and Samhita Alla
domain
flyte