Skip to footer

Architecture

Productionizing Distributed XGBoost to Train Deep Tree Models with Large Data Sets at Uber

We share technical challenges and lessons learned while productionizing and scaling XGBoost to train distributed gradient boosted algorithms at Uber.

Optimizing Observability with Jaeger, M3, and XYS at Uber

Uber’s observability engineers present their work on distributed tracing (Jaeger), sampling (XYS), and metrics processing (M3).

Introducing Menu Maker: Uber Eats’ New Menu Management Tool

To simplify the Uber Eats experience for our restaurant-partners, we built Menu Maker, a web-based tool for seamlessly managing menus on the Uber Eats app.

Evolving Michelangelo Model Representation for Flexibility at Scale

To accommodate additional ML use cases, Uber evolved Michelangelo's application of the Apache Spark MLlib library for greater flexibility and extensibility.
Uber Freight App

Building the New Uber Freight App as Lists of Modular, Reusable Components

We redesigned the Uber Freight app with RIBs, our open source plugin architecture, to enable quicker feature rollouts and an improved user experience.
word cloud

Less is More: Engineering Data Warehouse Efficiency with Minimalist Design

Data science helps Uber determine which tables in a database should be off-boarded to another source to maximize the efficiency of our data warehouse.

Migrating Functionality Between Large-scale Production Systems Seamlessly

With zero downtime, Uber's Payments Engineering team embarked on a migration that would allow authorization hold logic to be written once and used across existing and future payments products.

Employing QUIC Protocol to Optimize Uber’s App Performance

Implementing QUIC protocol against TCP over cellular networks on our apps led to a reduction of 10-30 percent in tail-end latencies for HTTP traffic.

Improving the User Experience with Uber’s Customer Obsession Ticket Routing Workflow and Orchestration Engine

Uber adopted workflow orchestration and Cadence, our open source orchestration engine, to better handle customer support ticket routing at scale.

Open Sourcing Peloton, Uber’s Unified Resource Scheduler

First introduced by Uber in November 2018, Peloton manages resources across large-scale, distinct workloads, combining separate compute clusters.

Using Machine Learning to Ensure the Capacity Safety of Individual Microservices

Uber leveraged machine learning to design our capacity safety forecasting tooling with a special emphasis on calculating a quality of reliability score.

Managing Uber’s Data Workflows at Scale

In this article, we discuss Uber's journey toward a unified, multi-tenant, and scalable data workflow management system.

Aarhus Engineering Internship: Building Aggregation Support for YQL, Uber’s Graph Query Language for Grail

Uber intern Lau Skorstengaard shares his experience working on YQL, the graph query language for our in-house infrastructure state aggregation platform.

The Billion Data Point Challenge: Building a Query Engine for High Cardinality Time Series...

Part of Uber's open source M3 metrics system, our query engine can support real-time, large-scale computation and multiple query languages.

Observability at Scale: Building Uber’s Alerting Ecosystem

Uber’s Observability team built a robust, scalable metrics and alerting pipeline to detect, mitigate, and notify engineers of issues as they occur.

Peloton: Uber’s Unified Resource Scheduler for Diverse Cluster Workloads

Uber developed Peloton to help us balance resource use, elastically share resources, and plan for future capacity needs.

Uber’s Big Data Platform: 100+ Petabytes with Minute Latency

Responsible for cleaning, storing, and serving over 100 petabytes of analytical data, Uber's Hadoop platform ensures data reliability, scalability, and ease-of-use with minimal latency.

Maximizing Process Performance with Maze, Uber’s Funnel Visualization Platform

Uber developed Maze, our funnel visualization platform, to identify possible UX bottlenecks and provide insight into the various ways riders and drivers interact with our platform.

M3: Uber’s Open Source, Large-scale Metrics Platform for Prometheus

M3, Uber's open source metrics platform for Prometheus, facilitates scalable and configurable multi-tenant storage for large-scale metrics.

Databook: Turning Big Data into Knowledge with Metadata at Uber

Databook, Uber's in-house platform for surfacing and exploring contextual metadata, makes dataset discovery and exploration easier for teams across the company.

Popular Articles