Staff Software Engineer (Developer Automation, Bazel)

• Build time is a tax every engineer pays on every change, and our stack collects it at a high rate. An MLIR-based compiler, Rust and C++ runtimes, and a large Python ML codebase have to produce identical results on a laptop, in CI, and on a machine with an accelerator in it. You will lead d-Matrix’s Bazel build-out. That means finishing the in-progress migration and designing the distributed build infrastructure that doesn’t exist yet
• You will own the build graph, toolchains, caching and remote execution, and the conventions every engineering team builds against. You will partner closely with development groups and with the engineers who run our runner fleet and cloud environments. When build and test time drops, every group in the company moves faster
• Complete the Bazel migration, including the hardest remaining components
• Migrate the test suite to Bazel, with correct tagging and sharding, and route hardware-dependent tests to executors that have the hardware
• Design, deploy, and operate remote execution in production
• Build hermetic toolchains and reproducible builds across developer machines, CI, and hardware-attached runners
• Raise cache hit rate to target and build trust in the cache, so no one runs a clean build “to be safe.”
• Establish a build-time baseline in your first quarter, then reduce median incremental build time and CI critical path against it
• Evolve the CI/CD system to keep pace with agent-assisted development
• Raise the standard of Bazel usage across teams by making the correct path the easy one and sharing best practices

Benefits

• Equity
• Health care
• Flexible time-off
• Paid paternity leave
• 401k retirement plan
• Remote & hybrid working model
• Work with world-leading engineers and researchers
• Peer coaching & development
• Free food and snacks
• Happy hours
• Team-building activities
• On-site fitness center- Strong knowledge of C++, Rust, Python, and container image builds
• Experience with hermetic, reproducible, and observable builds, including custom toolchains
• Persistence in debugging non-reproducible builds and unexplained cache misses to root cause
• A pragmatic bias toward incremental migration that keeps people productive
• Experience migrating a real codebase onto Bazel, including the long tail
• Hands-on experience deploying and operating remote caching or remote execution, self-hosted or vendor-provided
• A track record of partnering with other teams to understand their needs and spread best practices
• Deep Bazel experience as an owner: custom rules, toolchains, and target graph design in a large repository
• Experience migrating test suites to Bazel
• Kubernetes-based executor pools
• Open source contributions to the Bazel ecosystem
• Embedded or firmware cross-compilation. Firmware is not in Bazel today, and you would help decide whether it should be