Uber built Michelangelo, our machine learning platform, in 2015. Three years later, we reflect our journey to scaling ML at Uber and lessons learned along the way.
Uber developed Michelangelo PyML to run identical copies of machine learning models locally in both real time experiments and large-scale offline prediction jobs.
Uber's Advanced Technologies Group introduces Petastorm, an open source data access library enabling training and evaluation of deep learning models directly from multi-terabyte datasets in Apache Parquet format.
Uber AI Labs introduces Visual Inspector for Neuroevolution (VINE), an open source interactive data visualization tool to help neuroevolution researchers better understand this family of algorithms.
Written in Haskell, Queryparser is Uber Engineering's open source tool for parsing and analyzing SQL queries that makes it easy to identify foreign-key relationships in large data warehouses.
Migrating our Schemaless sharding layer from Python to Go while in production demonstrated that it was possible for us to rewrite the frontend of a massive datastore with zero downtime.
Uber Engineering created Omphalos, our new backtesting framework, to enable efficient and reliable comparison of forecasting models across languages.
Up for the challenge of developing at unprecedented scale? First, learn what it takes to master the technical interview process at Uber.
Uber Engineering built Denial by DNS, our open source solution for preventing DoS by DNS outages, to facilitate more reliable experiences on Uber's apps, no matter how users choose to access them.
Pyro is an open source probabilistic programming language that unites modern deep learning with Bayesian modeling for a tool-first approach to AI.
Uber Engineering's fraud prevention team built the Mastermind rules engine to detect highly evolved forms of fraud at large scale in a fraction of a second.
Seemingly small inefficiencies are greatly magnified as Uber's business scales. In this article we’ll explore design considerations and unique implementation characteristics of Pyflame, Uber Engineering's high-performance Python profiler implemented in C++.
A behind-the-scenes look at how Uber Engineering continues to develop our virtual onboarding funnel which enables hundreds of thousands of driver-partners to get on the road and start earning money with Uber.
The end of a two-part series on the tech stack that Uber Engineering uses to make transportation as reliable as running water, everywhere, for everyone, as of spring 2016.
Uber’s mission is transportation as reliable as running water, everywhere, for everyone. Here's the first of a two-part series on the tech stack that Uber Engineering uses to make this happen.
To show how a microservice is implemented in Uber Engineering's ecosystem, we look at the development of Tincup, our currency and exchange rate service.
Imagine you have to store data whose massive influx increases by the hour. Your first priority, after making sure you can easily add storage capacity, is to try and reduce the data’s footprint to save space. But how? This is the story of Uber Engineering’s comprehensive encoding protocol and compression algorithm test and how this discipline saved space in our Schemaless datastores.