TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A Pragmatic Engineer interview with Uber Distinguished Engineer Uday Kiran Medisetty details how the company’s developer infrastructure supports AI coding agents. Uber’s setup includes a model gateway, tools for switching coding harnesses, spending controls and human review of all pull requests.
The Pragmatic Engineer has published an interview with Uber Distinguished Engineer Uday Kiran Medisetty describing how the company’s internal developer infrastructure is being adapted for AI coding agents. The account outlines a model gateway, spending controls and agent-support tools, while stating that engineers still review every pull request.
Medisetty, described by The Pragmatic Engineer as one of Uber’s longest-tenured engineers and a leader in its work on agentic AI, discussed tools used by engineers to connect AI assistants with company code, internal services and engineering knowledge. The report names more than 20 internal tools and highlights DevPods, Code Inbox, Slate and Wisdom, although the supplied material does not explain each one’s function in detail.
Uber’s Model Gateway, introduced by its platform team in 2023, is intended to let AI development remain independent of a single model provider or harness. The report says it supports models from Anthropic and OpenAI, as well as open models Uber runs. Security middleware called AI Guard checks requests and responses; the platform also tracks spending by user and keeps audit logs and session traces.
Uber’s AIFX wrapper supports Claude Code, Codex and OpenCode, carrying settings, skills and plugins between them, according to the report. Engineers receive spending tiers and can seek a manager’s approval for more monthly usage. A plugin called CostCheck offers advice on costs. The source says Uber reduced token costs by about 50% in recent months, but does not provide the measurement period, baseline or methodology.
How Uber Governs AI Coding
The interview describes infrastructure choices that affect how AI tools are deployed across a large engineering organization. A shared gateway gives Uber a way to route work among different model providers while adding security checks, cost tracking and audit records. Supporting multiple coding harnesses may also let engineers change tools without rebuilding their individual configurations.
The approach reflects a practical tension: giving engineers access to AI assistance while managing spending and keeping people accountable for code changes. Uber says all pull requests remain subject to engineer review. The report also says the company has built review tools to help reviewers prioritize their attention, amid what Medisetty describes as a growing review bottleneck. The material does not establish that AI tools have improved productivity or code quality.
As an affiliate, we earn on qualifying purchases.
Infrastructure Built for Uber’s Scale
The interview places Uber’s AI systems within a longer history of in-house engineering tools. The report says Uber had two services when Medisetty joined in 2015, and thousands by 2019. It says the service count in 2026 will be about the same as in 2019, despite the business being about 10 times larger than it was then; the source does not give a precise counting method.
The company built systems including uOwn for service ownership, M3DB for metrics, Cadence for durable workflows and Morpheus for feature flags and experimentation. The report says Uber built these tools because available commercial products could not meet its needs or did not exist at the time. It also describes cities as a core operational concept that shapes services, deployments and cloud infrastructure, with each city pinned to a cloud region and able to fail over to another.
“All PRs are still reviewed by engineers.”
— The Pragmatic Engineer, summarizing the interview with Uday Kiran Medisetty
developer infrastructure security middleware
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limits of the Reported Results
The supplied report does not give independent measurements of how AI tools affect engineering speed, defect rates or review quality. Its claim that Uber reduced token costs by about 50% is not accompanied by a time window, comparison baseline or calculation method, so it cannot be evaluated as a like-for-like change from the material provided.
Some tools are named but not explained, and the source excerpt ends during its discussion of code-review features. It mentions AI review, screenshots and videos where applicable, and risk scores, but does not provide full details of how these inputs are generated or used. The interview is an account of Uber’s own systems, not an independent audit of their performance.
cost management tools for AI development
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Further Details on Review Tools
The Pragmatic Engineer episode is available on YouTube, Apple and Spotify, with a transcript and timestamps listed on its page. Those materials are the next source for fuller detail on how Uber’s review tools work, how the company measures productivity and how it sets AI spending tiers. The supplied report does not announce a new product launch or provide a schedule for additional infrastructure changes.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the news about Uber’s AI infrastructure?
The Pragmatic Engineer published an interview with Uber Distinguished Engineer Uday Kiran Medisetty about internal developer tools and infrastructure that support AI coding agents.
Which AI models and coding tools does Uber support?
The report says Uber’s Model Gateway supports Anthropic, OpenAI and open models run by Uber. Its AIFX wrapper supports Claude Code, Codex and OpenCode.
Does AI replace engineer review of Uber code?
No. The report says all pull requests are still reviewed by engineers. AI review and other tools are described as aids to the review process.
How does Uber manage AI spending?
Engineers receive spending tiers, and managers can approve additional monthly spending when a limit is reached. The report also describes CostCheck, a plugin that offers cost advice across sessions.
How much has Uber reduced token costs?
The source reports a reduction of about 50% in recent months. It does not state the measurement window, baseline or method, so the figure’s comparison basis is unclear.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
