About the Role
What we do
Delphi, Uber's Lever Efficiency team, builds the system Uber uses to recommend how to invest billions of dollars per year across its marketplace supply and demand levers: courier pricing (supply), courier incentives (supply), delivery and service fees (demand), user promotions (demand), etc. Starting with Uber's Delivery Marketplace business, we turn experiment results into a measure of each lever's incremental value in a common currency and publish it in a dashboard that refreshes automatically. We're also building a budget allocation tool that recommends how to move budget between levers. We work day to day with data scientists, product, operations, and finance.
Why it matters
Today each lever spends against its own budget target, even after it starts hitting diminishing returns. A trusted, comparable measure of what each lever returns lets Uber move money to where it does the most good: more reliable deliveries for customers, more earning opportunities for couriers, and more long-term growth for every dollar spent. The hard part is making the numbers trustworthy enough that Operations and Finance leaders act on them. Then the recommendations have to be proven with well-designed experiments. That's the system you'll help build.
Basic Qualifications
- Bachelor's or Master’s degree in Computer Science or a related technical field, or equivalent experience in an engineering related bootcamp
- 2+ years of experience in professional software development experience, developing and deploying systems in production environments
- Proven skills writing efficient, clean code in Go, Java, or Python
- Ability to debug and fix software bugs in a production scale environment
Experience with software engineering best practices
Strong grasp of distributed-systems fundamentals: caching, sharding, consensus, fault tolerance, observability, and incident response
Excellent written and verbal communication skills, including the ability to document architectures and algorithms
Preferred Qualifications
Advanced coursework or industry background in statistics, causal inference, or experimental design.
Direct ownership of analysis workflows (A/B, quasi-experimental, marketplace simulations) at scale.
Python, Pandas, PySpark
Data Pipelines (e.g. Spark, Scala)
Jupyter Notebooks
Fullstack development experience (BE, FE, Data Pipelines), this project involves a full suite of development surfaces across UI, backend, and data.
Experience developing internal tooling used by Ops, Engineers, Finance, Scientists, etc.
What the Candidate Will Do
Build and own data pipelines (Spark, Hive, scheduled workflows) that recompute lever-efficiency estimates from experiment and marketplace data every week. Include data-quality checks, freshness alerts and backfills.
Build backend services and APIs in Go that store and serve efficiency estimates and budget recommendations.
Ship full-stack features in the web app: interactive dashboards and a budget allocation tool used by Operations, Finance, Product, and Science.
Work with data scientists to turn notebook prototypes into production code. This includes confidence intervals, corrections for network effects, forecasts, and an optimizer that moves budget between levers within real constraints, such as fixed total spend and capped weekly changes.
Work with Uber's experimentation platform to run experiments that test budget moves (switchbacks, geo experiments, synthetic control), and automate reading their results into the dashboard.
Take features from design to launch: write design docs, review code, test, monitor production, and take part in on-call.
Document methodology and system design so partners can trust and reuse the numbers.
For San Francisco, CA-based roles: The base salary range for this role is USD $171,000 per year - USD $190,000 per year.
You will be eligible to participate in Uber's bonus program, and may be offered an equity award & other types of comp. All full-time employees are eligible to participate in a 401(k) plan. You will also be eligible for various benefits.
Job description provided by Uber.
View the original posting