KubeCon Amsterdam 2026: The Industrialization of ML - A Deep Dive into Uber’s AI Platform Architecture.
Uber's AI platform has evolved from fragmented, ad-hoc machine learning workflows into a highly scalable, Kubernetes-native system capable of handling millions of predictions per second. The company built Michelangelo to standardize ML workflows, but eventually migrated to Kubernetes and Ray to overcome scalability limitations for deep learning and large language models. This industrialized approach now underpins critical services across Uber’s global operations.
- ▪Uber runs over 1 million diverse workloads across 200 Kubernetes clusters and trains 20,000 models monthly.
- ▪Michelangelo, Uber’s internal ML platform, introduced a centralized feature store and standardized model deployment across offline, online, and library modes.
- ▪The shift from MADLJ to Kubernetes and Ray addressed resource fragmentation and improved scalability for GPU-intensive deep learning tasks.
Opening excerpt (first ~120 words) tap to expand
try { if(localStorage) { let currentUser = localStorage.getItem('current_user'); if (currentUser) { currentUser = JSON.parse(currentUser); if (currentUser.id === 3657823) { document.getElementById('article-show-container').classList.add('current-user-is-article-author'); } } } } catch (e) { console.error(e); } Soumia Posted on May 17 KubeCon Amsterdam 2026: The Industrialization of ML - A Deep Dive into Uber’s AI Platform Architecture. #kafka #kubecon #cloudnative #machinelearningops This article serves as a technical follow-up to our KubeCon 2026 coverage, providing a comprehensive deep dive into the architecture and evolution of Uber’s machine learning platform.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at DEV.to (Top).
Discussion
0 commentsMore from DEV.to (Top)
