In a world increasingly driven by software and digital services, the efficiency of a company’s developers has become a critical determinant of its success. As businesses scale their operations, they face the complex challenge of managing vast codebases and streamlining the workflows of thousands of engineers.
To address this, a new wave of innovation is emerging, where companies are strategically implementing AI to augment their developer teams. This is not just about adopting a new tool; it’s about fundamentally rethinking how software is built, tested, and maintained.
The focus has shifted from replacing human effort to empowering it, using AI to automate repetitive, time-consuming tasks and free up developers to focus on creative problem-solving and innovation. This strategic integration of AI into the software development lifecycle is quickly becoming a core pillar for modern, high-performing tech organizations.
Uber’s AI Developer Tool Strategy
Uber, a colossal company orchestrating 33 million trips daily across 15,000 cities, operates on a massive codebase comprising hundreds of millions of lines of code. The developer platform’s crucial mission is to ensure this intricate system runs smoothly and, perhaps more challenging, to keep 5,000 developers happy. This involves tackling developer toil in everyday tasks like writing tests or reviewing code, areas ripe for innovation and efficiency. To address this, Uber has strategically leveraged AI developer tools, primarily built using LangGraph.
Uber’s approach to AI dev tools rests on three core pillars:
- Product/Bet Areas: The team focuses on specific developer workflows like test writing and code review, aiming to eliminate toil and accelerate processes where they can make the most impact.
- Cross-Cutting Primitives: They develop foundational AI technologies, abstractions, and frameworks, such as Langfect, an opinionated framework that wraps LangGraph and LangChain. This allows for faster and more scalable solution development across the organization.
- Intentional Tech Transfer: After building initial products, Uber deliberately identifies reusable components to reduce the barrier for solving subsequent problems. Langfect itself was born from this need, providing reusable nodes for agentic solutions, with LangGraph proving a perfect fit.
Revolutionizing Developer Workflows with AI Products
Uber has rolled out several impactful AI-powered tools:
- Validator: This IDE experience automatically flags best practice violations and security issues in code. Built as a LangGraph agent, Validator notifies engineers of problems, offering the choice to apply a precomputed fix or send the request to their IDE agentic assistant. It smartly combines feedback from LLMs with deterministic capabilities, like static linters, to discover and pre-compute fixes for lint issues. This results in thousands of fixed interactions daily, meeting developers in their natural IDE environment and providing actionable solutions.
- Autocover: Moving beyond validation, Autocover assists engineers in generating high-quality tests—building passing, coverage-raising, business-case, validated, and mutation-tested tests. Its primary goal is to save engineers time, enabling them to quickly generate tests and move on to new features. Autocover is powered by a collection of expert domain agents, notably including Validator. When invoked, it sets up the test environment, analyzes source code for business context, and then streams generated tests that are continually built and refined. Its graph incorporates specialized agents like the Scaffolder (prepares environment, identifies business cases), the Generator (devises new test cases), and the Executor (runs builds and tests, checks coverage). Autocover can perform 100 iterations of code generation and 100 executions simultaneously. Benchmarked against industry tools, it achieves two to three times more coverage in about half the time, largely due to its custom, bespoke knowledge built into its agents. This tool has raised developer platform coverage by approximately 10%, translating to 21,000 developer hours saved and thousands of tests generated monthly.
Proliferation of AI Tools Across Uber
The success of these primitives has led to a proliferation of AI tools:
- Uber Assistant Builder: An internal custom GPT store enabling chatbots steeped in Uber knowledge, such as the Security Scorebot, which detects security anti-patterns.
- Genie Adopt: A conversational AI integrated into Picasso (Uber’s internal workflow management platform) that understands workflow automations and provides feedback grounded in product truth.
- U review: A tool that flags code review comments and suggestions during the code review process to enforce quality before code merges, leveraging similar underlying tools as Validator and Autocover.
Ultimately, Uber’s journey highlights that building super-capable domain expert agents and composing them with deterministic sub-agents yields exceptional results and enables scalable development through reusability. The use of graph abstractions like LangGraph not only accelerates AI workloads but also improves the experience for non-AI interactions, proving mutually beneficial for traditional and agentic systems.
Key Takeaways
- Uber utilizes AI developer tools, built using LangGraph, to improve developer efficiency and happiness.
- Key AI tools include Validator (IDE code checking) and Autocover (automated test generation).
- These tools have resulted in significant time savings and increased code coverage for Uber’s developer platform.
- The company emphasizes reusable AI components and a strategic approach to technology transfer.
Join our community by subscribing to our Weekly Newsletter to stay updated on the latest AI updates and technologies, including the tips and how-to guides. (Also, follow us on Instagram (@inner_detail) for more updates in your feed).
(For more such interesting informational, technology and innovation stuffs, keep reading The Inner Detail).
