Responsibilities
- Design, develop, and maintain machine learning models and software for personalization, candidate generation, multi-objective ranking, and recommendation features.
- Build foundational ML infrastructure, automated model retraining pipelines, and near-real-time streaming feature pipelines using technologies such as Flink and Kafka.
- Own the ML software lifecycle from design documents and feature extraction pipelines through inference monitoring, performance optimization, and production incident resolution.
- Collaborate with Product, Data Science, Platform, and Ops teams to translate ambiguous requirements into scalable ML system architectures.
- Establish code and model architecture reviews, offline validation, online A/B testing, model monitoring, feature-quality tracking, and model-performance monitoring.
- Optimize inference throughput and monitor latency SLAs, feature drift, label leakage, and production decisioning reliability.
Requirements
- At least 2 years of professional software engineering experience in machine learning.
- Hands-on experience with deep learning and machine learning frameworks such as PyTorch and TensorFlow.
- Strong knowledge of feature engineering, loss-function design, hyperparameter optimization, and offline evaluation metrics.
- Strong proficiency in Python and one or more object-oriented or functional languages such as Go, Java, or C++.
- Bachelor’s or Master’s degree, or equivalent experience, in Computer Science, Machine Learning, Data Science, or a related technical field.
- Experience writing unit and integration tests, verifying model functionality, and implementing basic system monitoring.
- Preferred experience designing, deploying, and maintaining production recommendation, ranking, or personalization systems.
- Preferred experience with central feature stores such as Palette, embeddings, real-time or near-real-time feature engineering pipelines, model monitoring, feature skew and drift detection, automated offline evaluation, online experimentation, Spark, Flink, Kafka, and distributed training.
Tech Stack
Categories
About Uber
We are Uber. The go-getters. The kind of people who are relentless about our mission to help people go anywhere and get anything and earn their way. Movement is what we power. It’s our lifeblood. It runs through our veins. It’s what gets us out of bed each morning. It pushes us to constantly reimagine how we can move better. For you. For all the places you want to go. For all the things you want to get. For all the ways you want to earn. Across the entire world. In real time. At the incredible speed of now. The idea for Uber was born on a snowy night in Paris in 2008, and ever since then our DNA of reimagination and reinvention carries on. We’ve grown into a global platform powering flexible earnings and the movement of people and things in ever expanding ways. We’ve gone from connecting rides on 4 wheels to 2 wheels to 18-wheel freight deliveries. From takeout meals to daily essentials to prescription drugs to just about anything you need at any time and earning your way. From drivers with background checks to real-time verification, safety is a top priority every single day. At Uber, the pursuit of reimagination is never finished, never stops, and is always just beginning.