Uber is one of those organizations that rely heavily on data. Each day, millions of journeys take place in 700 cities internationally, generating data on traffic, preferred routes, estimated times of arrival/delivery, drop-off locations, and more that allows Uber to deliver a smooth riding experience to its customers.
With access to the rich dataset coming from the cabs, drivers, and customers, Uber has been investing in machine learning and artificial intelligence to enhance its business. Uber AI Labs consists of ML researchers and practitioners that translate the advantages of the state-of-the-art machine learning methods and advancements to Uber’s core business. From computer vision to conversational AI to sensing and perception, Uber has efficiently infused ML and AI into its ride-sharing platform.
Since 2017, Uber has been sharing the best practices of building, deploying, and managing machine learning models. Some of the internal instruments and frameworks used at Uber are built on high of common open source projects such as Spark, HDFS, Scikit-learn, NumPy, Pandas, TensorFlow, and XGBoost.
Michelangelo is a machine learning platform that standardized the workflows and instruments across teams through an end-to-end system. It enabled builders and data scientists across the company to build and operate machine learning systems at scale.