About the job
The mission of the Surge team is to maintain overall marketplace reliability by balancing supply/demand in real-time through dynamic pricing. We build scalable real-time systems to understand the state of the market, forecast future demand, make predictions using ML models, solve network optimization programs, and eventually make pricing decisions for each rider session. Surge plays a critical role in service of Uber’s mission to make transport accessible. We generate billions of dollars in annual gross bookings for the company by optimizing network efficiency and make a significant contribution to driver earnings. In addition to pricing, the signals we generate are some of the most important features used in practically every optimization/ML system across Uber. Although we are a backend team, what we do has an outsized impact on our riders because prices and reliability are two of the most important elements of customer experience.
Responsibilities
Build and train machine learning models with sparse data
Design experiments and use a variety of techniques for building causal models
Be a thought leader and help define roadmaps across multiple rider pricing teams
Qualifications
Minimum
PhD in relevant fields (CS, Stats, Economics, Econometrics, etc.) with a focus on Machine Learning.
4+ years of experience in an ML role with an emphasis on data and experiment driven model development.
Expertise with Causal Inference, DML, etc...
Expertise in deep learning and optimization algorithms.
Experience with ML frameworks such as PyTorch and TensorFlow.
Experience building and productionizing innovative end-to-end Machine Learning systems.
Proficiency in one or more coding languages such as Python, Java, Go, or C++.
Strong communication skills and can work effectively with cross-functional partners.
Strong sense of ownership and tenacity toward hard machine-learning projects.
Preferred
Academic background in Economics or Econometrics
Experience in combining observational data with experimental data for building causal models.
Experience designing embeddings and combining structural models and regularization techniques for dealing with sparsity.
Experience building elasticity models and user behavioral models
Proven track record in conducting experiments and tracking models in high-complexity environments.