What we’re about
Check out the people and code behind Uber Tech in Amsterdam at our latest tech meetups and explore our work across Uber, Uber Eats, and more.
Our work:
Eng blog → http://eng.uber.com/
Develop API → https://developer.uber.com/
Open Source → http://uber.github.io/
Our team:
Careers blog → https://www.uber.com/blog/careers/
Twitter → https://twitter.com/UberEng
LinkedIn → https://www.linkedin.com/showcase/uberengineering/
Join us:
Open jobs in Amsterdam → https://www.uber.com/careers/locations/amsterdam/
Tour our Amsterdam office → https://www.uber.com/blog/inside-uber-amsterdam
Upcoming events (1)
See all- From Predictive to Agentic: How Michelangelo Accelerates Uber's AI JourneyBurgerweeshuispad 301, Amsterdam
**Registration Closed - Full Capacity**
AI at Uber is evolving rapidly, from predictive models powering critical business operations to agentic systems that drive decision-making and automation at scale. At the heart of this transformation is Michelangelo, Uber’s in-house ML platform, enabling teams to build, deploy, and scale AI solutions across the company.
Join us for an evening of talks and a fireside chat with Viv Keswani, Senior Director - ML Platforms, exploring how Michelangelo powers AI today and where it’s headed next.
📅 Date: October 14, 2025
📍 Location: Uber HQ, Tripolis Park, AmsterdamSummary of the talks:
Uber's Multi-Agent Platform by Rush Tehrani
In this presentation, we will introduce Uber’s Agent Platform, a unified platform to accelerate the development, deployment, evaluation and monitoring of production-ready AI agents at Uber. The platform offers two complementary approaches to meet diverse user needs: Uber Agent Builder, a no-code solution that empowers anyone to build and launch agents with ease, and AgentFx, a flexible, code-first SDK built for engineers who want deeper control and customization.
Scaling Sustainability: How Uber Eats Classifies Packaging with Multimodal LLMs by Felipe Masculo
Uber is committed to helping merchants adopt more sustainable packaging. However, manually assessing packaging for millions of merchants is slow, expensive, and doesn't scale. This talk presents our solution: a highly scalable system that leverages multimodal LLMs to automatically classify packaging types and materials from photos. We'll do a deep dive into our multi-stage pipeline, discuss the trade-offs that led us to use LLMs over traditional computer vision models, and share our evaluation framework and key results.
Agenda:
5:30PM - Doors Open
6:00PM - Intro to Uber Tech Amsterdam and Michaelangelo
6:15PM - Uber's Multi-Agent Platform
6:45PM - Scaling Sustainability: How Uber Eats Classifies Packaging with Multimodal LLMs
7:15PM - Fireside Chat with Viv Keswani and Melda Salhab
7:45PM - Q&A
8:00PM - Networking including snacks and drinks
9:00PM - Doors Close
There is limited seating so please only attend if you have RSVP'ed.
See you all there!
Not open
