engineering
Posted May 25Software Engineer, Build Systems / CI
at openai
San Francisco, United StatesOn-site
Responsibilities
- - Design and maintain Starlark rules, macros, toolchains, and integrations that make builds reproducible, hermetic, and easy for product teams to adopt.
- - Build systems that reduce unnecessary CI work through affected-target detection, dependency graph analysis, test selection, caching, batching, and smarter scheduling.
- - Own the reliability of the systems you build, including participating in an on-call rotation for critical developer infrastructure.
Requirements
- About the Role The Engineering Acceleration team builds and operates the foundational systems that engineers use to build, test, and ship ChatGPT, the API, and OpenAI's infrastructure.
- You will work on the systems that determine how quickly and confidently engineers can move: Bazel-based builds, Buildkite pipelines, test selection, remote caching and execution, CI observability, and tooling that helps engineers understand and fix failures quickly.
- - Operate and optimize build infrastructure across Docker/OCI images, Kubernetes-based runners, cloud resources, and remote cache/execution systems.
- - Use modern AI tools to rethink CI failure analysis, flaky test debugging, PR triage, automatic remediation, and developer-facing explanations.
- Technologies Commonly Used In This Environment Include - Bazel and Starlark for build and test workflows - Buildkite for CI orchestration - Docker and OCI images for build and runtime packaging - Kubernetes for CI runners and infrastructure orchestration - Python, Go, TypeScript, Rust, C++, and other languages in a large monorepo - Terraform for infrastructure as code - Remote caching, remote execution, artifact storage, and build telemetry systems - Postgres, Kafka, and internal services used to power
- experience building infrastructure or tooling for developers. - Have hands-on
- experience with Bazel, Buck, Pants, Gradle, or similar build systems, and understand the tradeoffs of hermetic builds, dependency graphs, caching, sandboxing, and remote execution. - Have built or operated CI systems at scale, especially in environments where build time, queue time, test flakiness, and developer trust materially affect engineering velocity. - Are comfortable writing production software for internal platforms, not just configuring tools.
- experience and have empathy for the small sources of friction that slow teams down or create operational toil. - Are pragmatic about platform adoption: you know how to build paved paths that teams want to use because they are faster, clearer, and more reliable. - Communicate clearly across teams and can turn ambiguous productivity problems into concrete technical plans. - Are excited to apply AI to developer infrastructure in ways that make engineers faster without weakening quality, reliability, or safety.
- About OpenAI OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence
Additional details
- This role sits at the intersection of developer productivity, build systems, distributed infrastructure, and software quality.
- Our mission is to make OpenAI one of the most productive engineering organizations in the world while preserving a high bar for correctness, reliability, and safety.
- The best version of this work is invisible when it succeeds: builds are fast, tests are trusted, CI failures are understandable, and engineers can focus on shipping useful systems instead of fighting infrastructure.
- In This Role, You Will - Own and evolve Bazel-based build and test workflows across a large, polyglot monorepo.
- - Improve CI performance and reliability across Buildkite pipelines, including queue time, build time, cache hit rates, test sharding, retry behavior, and flake isolation.
- - Improve local development workflows so engineers can reproduce CI behavior, debug build failures, and iterate quickly without learning every detail of the build stack.
- - Instrument build and CI systems with metrics, logs, traces, dashboards, and analytics so we can measure speed, reliability, cost, and developer impact.
- - Partner directly with product, infrastructure, and research engineering teams to understand pain points, onboard projects, debug hard build issues, and remove systemic bottlenecks.
- We expect this role to involve code, design, debugging, operations, and long-term ownership. - Can debug distributed build and CI failures across source control, dependency management, containers, runners, remote caches, test frameworks, and service infrastructure. - Care deeply about developer