Build Systems Engineer (Bazel)

MatX•Mountain View, CA

22h•Hybrid

About The Position

MatX builds custom AI accelerator silicon. The build system is the backbone: it wires together RTL, a stack of commercial EDA tools (simulation, synthesis, place-and-route, lint/CDC), and a Rust/Python software stack into one hermetic, reproducible pipeline. We run on Bazel with bzlmod, RBE, custom rules, and a small but tight set of platform-level abstractions. You'll join a small group that owns the build graph, the toolchains, the rules that wrap each EDA tool, and the CI infrastructure that keeps thousands of targets green.

Requirements

Deep build-system fluency. Rules, providers (or equivalent), aspects, toolchains, platforms, configuration/select, transitions, query. Ability to read a build-system file — .bzl, Buck2 BUCK, Shake Rules.hs, whatever — and predict what its action graph will look like.
Bazel-native is a plus; we'll trade six weeks of Starlark ramp for the right taste.
Build-system fluency, not Bazel-keyword-matching. Equivalent work in Buck2, Shake/Hadrian, Pants, Nix, or a homegrown Blaze-shaped system is applicable. Be honest about ramp on Starlark and bzlmod.
bzlmod / MODULE.bazel. Module extensions, lockfile management, vendoring third-party deps cleanly
Remote execution. RBE, Buck2 RE, BuildBuddy, BuildBarn, your own — they all teach the same lessons. Cache-key debugging, Build without the Bytes, diagnosing "works locally, fails remote." Experience owning one end-to-end.
Comfort in Rust / Python / shell / Starlark. Ability to read all four in any given week.

Nice To Haves

Build graph is the source of truth. "If two things must stay in sync, make one depend on the other." Allergic to parallel lists in workflow YAML, Python arrays, and .bzl dicts that drift
Don't parse what you can generate. If a tool has the structured data internally, have it write structured output. Parsing human-readable reports is a temporary bridge, not a design
Split build from check. A rule that produces artifacts always succeeds; a separate _test target gates on quality. Empty dashboards because the build broke are unacceptable
Let Bazel parallelize, not the orchestrator. One bazel build --keep_going over N matrix jobs that each warm up Bazel
Encode execution constraints in the rule, not the invocation. No README accumulating per-tool --strategy=..., --remote_download_outputs=..., --sandbox_debug incantations. execution_requirements belongs on the action
Compose at the boundary. Dev and prod differ only in where data comes from — share the components, swap the source. No parallel dev-only servers
Smart targets over comprehensions. A direct rule call per file beats a list-comprehension factory plus three parallel side dicts
Sort keys for determinism. The bytes of an emitted file are a function of semantic content, not iteration order
Internal vs external API. Collapsing N files to 1 is a free improvement when you own both ends. It's a breaking change when someone else's scripts source those files by name
Migrated a non-trivial codebase between major build-system versions or paradigms — WORKSPACE → bzlmod, Buck1 → Buck2, Make → Shake, or similar. Bonus if you led the bzlmod / MODULE.bazel housekeeping
Written rules that wrap closed-source tools (EDA, ML, codegen) and made them hermetic and remote-executable — in Bazel, Buck2, Shake, or otherwise. The craft is the hermetic wrapper, not the syntax
Owned a remote-execution deployment end-to-end — Bazel RBE, Buck2 RE, BuildBarn, BuildBuddy, in-house — including diagnosing cache poisoning, toolchain drift, or remote-vs-local divergence

Responsibilities

New EDA tool integrations. Wrap a closed-source tool in a hermetic Bazel rule with proper providers, runfiles, and execution constraints. Add a new front-end stage to an existing toolchain; add a rule for test variants that share configuration; wire a third-party generator into our verilog graph as a first-class dep
Bazel version migrations. Lead upgrades (8.x → 9.x) and the bzlmod/MODULE.bazel housekeeping that comes with them
Hermeticity work. Hunt down the implicit assumptions: system Python, system gcc, leaked /usr/bin deps, host-state in tests. Replace them with hermetic toolchains and tracked inputs
Refactors that delete code. Rewrite a fragmented family of test macros in terms of one shared rule. Remove a homegrown wrapper rule once upstream covers the case. Extract a common aspect helper used by three places that duplicated it. The good PRs net negative
Build performance. Persistent workers for slow tools, RBE configs, action graph hygiene, cache-key debugging when something silently rebuilds
CI infrastructure. GitHub Actions self-hosted runners on GCE COS, Buildbarn workers, monitoring, rolling upgrades
Teach the rest of the team Bazel. Half the company writes RTL or Rust, not Starlark. Good rules let them stay in their domain. Good docstrings (and stardoc) keep them self-serve
Work tightly with at least one of us. Most non-trivial changes are pair-designed before code. Fast feedback loops, whiteboard sessions, no async-only collaboration
Lean on AI, but stay persnickety. Use Claude Code and similar tools heavily — for prototypes, refactors, scripts, even rule scaffolding. Reject most of what they produce on the first pass. Steer the model hard toward your taste, push back on the easy answer, and review every line you commit as if you wrote it. Auto-generated PRs that pass tests but miss the point are not what we want

Benefits

A Stake in our success A cash/equity mix that fits your needs, and option to do early exercise
Health & Wellness Company subsidized Health, Dental, Vision, and Life insurance; Pre-tax Health Savings Accounts with generous company contribution (even if you don’t)
Time To Recharge 4 weeks paid time off (accrued), 12 company holidays, and 3 weeks remote/flexible work per year
Support to Parents Up to 12 weeks of paid parental leave, regardless of your path to parenthood
Learning & Development $1,500 yearly towards your professional development e.g. conferences, courses, and other learning opportunities
Team Connection Team Lunches, quarterly off-sites, and regular town halls
Financial Wellbeing 401K and/or Roth IRA, with 5% company contribution, even if you don’t!
Flexible Spending Accounts Pre-tax spend accounts for medical, dental/vision, dependent care, parking, and transit expenses
Commute On Us For those commuting up to 1 hour, put your rideshare cost on our company card and reclaim the drive-time to get work done!
MatX E[x]tras $50 per month to use on the perks you care about most
Remote Perks We work remotely Monday & Friday, supported by home-tech setup, and remote wifi expense reimbursement