Data Scientist - Forecasting Uber
San Francisco, CA

Job Description

Uber Overview

At Uber, we ignite opportunity by setting the world in motion. We take on big problems to help drivers, riders, delivery partners, and eaters get moving in more than 600 cities around the world.

We welcome people from all backgrounds who seek the opportunity to help build a future where everyone and everything can move independently. If you have the curiosity, passion, and collaborative spirit, work with us, and let's move the world forward, together.

Job Description

About the role

Uber's Marketplace team is looking for experienced, passionate data scientists to join a functional team devoted to supply, demand, trips and gross booking forecasting. Forward projections of drivers and riders, across varying levels of spatial granularity and multiple timescales, are crucial building blocks for Uber's pricing and supply positioning systems. Creating and updating these forecasts in response to a rapidly changing world is a fascinating and challenging problem, which must be solved in real time leveraging enormous data sets.

What You'll Do / What You'll Need / Bonus Points / About the Team

What You'll Need

* Excellent educational background in machine learning, statistics, computer science, applied math, signal processing, electronic trading, economics, operations research, genomics, fluid dynamics and computational modeling or a related field. Ph.D. or Masters degree preferred.
* Entrepreneurial mindset. Everywhere you go, you can't help but mobilize people, create things, solve problems, roll up your sleeves, collaborate, go above and beyond. You are an insatiable doer and motivator of others.
* Excellent execution and organization. This team works with engineers and product leads at the forefront of the development cycle. To excel in this role, you should be comfortable executing with little oversight and be able to adapt to problems quickly.
* Experience with common analysis tools: Python, R, and SQL. Demonstrable abilities with code and programming concepts.

Nice to Have

* 3+ years industry experience in time series modeling or machine learning, with significant personal experience as a technical contributor.
* Experience working with large data sets; experience with spatial data.

What You'll Do

* Construct time series and machine learning models to forecast fundamental business quantities such as supply, demand trips and gross booking.
* Extract data from warehouses, back-test models, and compare model performance and communicate results.
* Create dashboards and reports to regularly communicate results and monitor key metrics.
* Design experiments and interpret results to draw detailed and actionable conclusions. Perform time-series analyses, hypothesis testing, and causal analyses to statistically assess relative impact and extract trends.
* Collaborate with cross-functional teams, including product, engineering, operations, and marketing.

About the Team

We are building the brain that powers Uber. The Marketplace Team (https://www.uber.com/marketplace) is the foundation for Uber's network where riders and drivers come together at extraordinary scale. In Marketplace, we build and scale the real-time systems that power intelligent dispatch, dynamic pricing, driver positioning, experimentation and real-time forecasting.

We are solving some of the hardest data processing, science, and engineering problems related to Uber's real time marketplace. The problems include ingesting, transforming, and analyzing massive amounts of data, both structured and unstructured, by building distributed systems and algorithms to power the real-time Marketplace decisions and strategic business moves.

Data Scientists on Marketplace combine clear product vision, deep technical skills and powerful data analysis and modeling to improve the algorithms that run Uber's vast worldwide marketplace every minute of the day. Marketplace forecasting stands at the upstream of the marketplace optimization flow and generates time series and machine learning models for quantities like supply, demand, marketplace balance trips and gross bookings. These forecasts provide a forward view of marketplace and empower downstream optimization algorithms such as pricing and incentive allocation in a real-time manner. These forecasts aim to model the world and proactively capture calendar patterns and shocks (holidays, events, weather and more) using a combination of classical time series and statistical and machine learning approaches in a scalable way.