Wecome to a new edition of the MLOps Community Meetups in Amsterdam! This time, Uber is hosting our next event and we couldn't be more excited.

📅 Agenda

17:30 - 18:15: Walk-in 🚶‍♂️
18:15 - 18:45: Understanding LLM Quantization and Running Models on Your Laptop - Sarthak Anand
18:45 - 19:15: Trusting AI to Modernize Software at Scale - Tim te Beek
19:15 - 19:25: Quiz w/ prizes + small break.
19:25 - 19:55: Scaling Mobile ML to Billions of Daily Inferences: What Breaks and How to Fix It - Henrique Vasconcelos
19:55 - 20:05: Lightning talks ⚡️
20:05 - 21:00: Networking + drinks and bites 🍹🍴

Talks:

#1. Understanding LLM Quantization and Running Models on Your Laptop - Sarthak Anand
Ever wondered how to run powerful AI models on your own computer? How do massive LLMs fit onto regular laptops? What’s the secret behind running AI without cloud services? This talk addresses these questions by exploring how quantization makes powerful language models run on everyday hardware. Learn the idea behind GGUF format, understand the trade-offs between model size and performance, and get practical guidance for deploying AI locally.

#2. Trusting AI to Modernize Software at Scale - Tim te Beek
Tim explores how machine learning and generative AI can make large-scale code migrations faster, safer, and more efficient. Drawing on lessons from building Moderne’s AI-driven transformation tools (built on OpenRewrite), he shares practical strategies for designing scalable workflows—covering code reuse, input validation, and result verification. The talk also demonstrates how orchestration frameworks like MCP and LangChain enable reliable, agent-driven automation across massive codebases.

#3. Scaling Mobile ML to Billions of Daily Inferences: What Breaks and How to Fix It - Henrique Vasconcelos
This presentation addresses critical challenges encountered during the a mobile model's rollout. It details the debug and investigation process of issues that appear when a model run on millions of separate devices. We go in depth detailing the investigation and eventual fixes for such problems, detailing Uber's engineering practices and mobile release cycle. In the end we look into the steps we're taking to try to prevent similar issues before they reach production.

Stick around for pizza, drinks, and great conversations with fellow AI and ML enthusiasts!

Want to chat with the rest? Join our Slack in the #Amsterdam channel!
https://join.slack.com/t/mlops-community/shared_invite/zt-36q0g9r83-qZLH7z2UA8~auwhfO7x1ZA

MLOps Community @ Uber