Data Infrastructure Engineer - Machine Learning Platform

2 days ago
US-CA-San Francisco
Engineering - Engineering
US-CA-San Francisco
Requisition Post Information* : External Company Name
Uber Technologies, Inc.
Requisition Post Information* : External Company URL

Uber Overview

About Uber


We’re changing the way people think about transportation. Not that long ago we were just an app to request premium black cars in a few metropolitan areas. Now we’re a part of the logistical fabric of more than 600 cities around the world. Whether it’s a ride, a sandwich, or a package, we use technology to give people what they want, when they want it.


For the people who drive with Uber, our app represents a flexible new way to earn money. For cities, we help strengthen local economies, improve access to transportation, and make streets safer.

And that’s just what we’re doing today. We’re thinking about the future, too. With teams working on autonomous trucking and self-driving cars, we’re in for the long haul. We’re reimagining how people and things move from one place to the next.

Job Description

About The Role


We are looking for experienced data infrastructure engineers to join our Machine Learning platform team. This group is perfect for those engineers looking to tackle the types of distributed data processing and machine learning challenges that are critical to Uber’s continued success.


What You’ll Do

  • Build infrastructure for batch and streaming data processing for model training and serving
  • Optimize data systems for throughput and efficiency
  • Build and Scale Uber’s distributed offline and online training infrastructure and algorithms

What You’ll Need

  • Experience with Java/Scala, Python, C++, or related languages/technologies are required 
  • Deep experience with big data or other distributed systems
  • Team player: You believe that you can achieve more on a team — that the whole is greater than the sum of its parts. You rely on others' candid feedback for continuous improvement
  • Clear Thinker: We are looking for engineers who can think clearly and can write code that reflects that clarity
  • Great Communicator: You can describe and discuss technical issues easily

Bonus Points

  • Previous experience with machine learning platforms, especially familiarity with modern ML packages (e.g. TensorFlow, Caffe, sklearn, MLLib, xgBoost)
  • Distributed/parallel learning algorithms
  • Online or continuous learning systems
  • Strong practical and theoretical understanding of distributed implementations of decision trees, linear models, optimization algos and scientific computing techniques
  • Strong practical and theoretical understanding of Deep Learning infrastructure on GPUs, tool kits, and model architectures

About the team


The Machine Learning Platform team builds the end-to-end systems and tools to enable teams around Uber to build and deploy machine learning solutions at scale. The platform is used by more than a dozen teams around Uber, including EATs, Map Services, Fraud, Marketplace, Finance, and ATG (autonomous cars).


The platform supports the full ML workflow:

  • Data management - unified online & offline systems for computing, storing, and sharing features for batch training and serving and online serving
  • Train models - distributed training of wide array of traditional ML, Deep Learning, time series forecasting, and bespoke algorithms
  • Evaluate models - version control and rich visualizations for trained models 
  • Deploy models - fully managed deployment of trained models across data centers to batch and online serving containers (deployment to cars and phones coming soon)
  • Make predictions - low latency, high throughput model serving 
  • Monitor predictions - monitor feature and prediction distributions and join predictions back to collected labels to provide ongoing monitoring of the accuracy of deployed models


Connect With Us!

Not ready? Connect with Uber to receive future communications about opportunities and general information about specific areas of Uber that interest you.