Back to calendar

Lunch Lecture Uber

Powering Machine Learning at Uber: Inside Michelangelo

Ever wondered how Uber helps match riders with drivers or recommends restaurants you might enjoy all in real time? 

Behind predictions like these is much more than an accurate model. Machine-learning systems must move reliably from experimentation to production, receive the right data, and serve millions of requests efficiently. That’s where Michelangelo—Uber’s machine-learning platform—comes in. 

As software engineers working on Michelangelo, we build infrastructure that helps teams develop, train, deploy, and operate machine-learning solutions. This gives us a unique perspective on the complete ML lifecycle: how models evolve from promising experiments into dependable production systems.

What we’ll cover

01. An introduction to Michelangelo and ML platform engineering 

  • An introduction to Michelangelo and how machine-learning models
    move from experimentation into production.

02. A closer look at Palette, Michelangelo’s Feature Store 

  • What features are and how models use them
  • Keeping feature data consistent between training and production
  • Making features reusable and reliable across ML applications
  • Operating feature infrastructure at Uber scale 

03. Glimpse on our journeys

  • Aleksandra Andrasz: How the Uber internship experience helped me feel empowered during my Master thesis and how it feels to be back.
  • Rana Afifi: From backend engineering on product-focused teams to ML platform infrastructure—and what engineering looks like across product and platform organizations 

Join us to discover the engineering foundations behind production machine learning—and how reliable infrastructure turns models into real-world impact.

Uber Logo

Register now