AI Platform Engineer, DevX

Brain Co.•San Francisco, CA

22d

About The Position

You'll join the DevX pod — the team that owns the build system, CI/CD, monorepo, and developer tooling that set the ceiling on how fast every other engineer at BrainCo can ship. This is a force-multiplier role: your work compounds as we add teams, services, and customers. BrainCo's monorepo spans Python, TypeScript, and Go, runs on Nx, and ships to production environments that include regulated and air-gapped customer deployments. As we grow from a handful of teams to dozens, the leverage of this role compounds. Your job is to discover, measure, and remove developer friction. You should be comfortable going from "this workflow is slow or confusing" to instrumented evidence, a shipped improvement, and clear before / after impact. Own the monorepo and build system Own the architecture and ongoing health of our Nx-based monorepo: project graph, dependency boundaries, language toolchains (pnpm, uv, Go), and code generation patterns. Make builds and tests fast, reliable, and incremental at scale. Design and operate CI/CD Build CI/CD systems with strong guarantees around correctness, reproducibility, and speed. Own pipeline architecture, caching, parallelism, secret flow, and release engineering for environments that include regulated and air-gapped customer deployments. Developer tooling and workflows Design internal CLIs, scaffolding, local dev stacks (Tilt, Docker, Kubernetes), and code-quality infrastructure (lint, typecheck, format, pre-commit). Own the developer-facing experience for deployment and environment workflows, while partnering with Infrastructure on the underlying cloud, Kubernetes, networking, and runtime substrate. Create paved paths the team actually wants to use — not guardrails that get bypassed. Treat developer tools as production systems Run the build, CI, and dev infra with real SLOs, observability, and on-call rigor. Internal downtime is downtime. Solve ambiguous productivity problems Identify what's slowing engineers down, instrument it, and fix it from first principles. Move from "X is slow" to a measured, deployed improvement with clear before/after numbers. Operate this like a real product End-to-end ownership: architecture, implementation, rollout, observability, on-call, and iteration based on internal user feedback. Partner closely with every engineering team. Drive adoption through quality, not mandates. Why This Role Matters You'll set the ceiling on how fast every other engineer at BrainCo can ship. As we grow from a handful of teams to dozens — building AI systems for governments and regulated industries — the platform that lets us move quickly without breaking what matters becomes one of the highest-leverage systems in the company.

Requirements

We're looking for engineers who treat developer experience as a product — engineers who reach for users, not just specs.
Developer Platform & Build Systems
Have 5+ years building developer-facing infrastructure: monorepo platforms, build systems, CI/CD, or internal tooling.
Have hands-on depth with at least one modern monorepo tool (Nx, Bazel, Pants, Buck, Turborepo) and the willingness to go deeper.
Have strong fluency with CI/CD systems at scale (GitHub Actions, Buildkite, CircleCI, or similar): matrix design, caching strategies, secret flow, parallelism.
Are fluent in at least two of: Go, Python, TypeScript, Rust, Bash.
Platform & Product Mindset
Have shipped developer platforms that real engineers adopted.
Have measured developer friction and shown clear before / after impact from tooling, CI, build, release, or workflow improvements.
Have strong intuition for developer experience, paved paths, and long-term maintainability.
Are comfortable owning the full lifecycle: design doc, MVP, hardening, adoption.
Production & Reliability
Have owned developer-facing systems with real uptime and operational responsibility.
Are familiar with observability tooling and incident response.

Nice To Haves

Built or significantly contributed to a remote build cache, custom build system, or hermetic build pipeline (Bazel, Nix, etc.).
Led a polyrepo → monorepo (or reverse) migration at scale.
Experience applying DORA / SPACE or other DX metrics in practice.
Background in release engineering, supply-chain security (SLSA, Sigstore), or artifact management.
Built internal PaaS, dev environments, or scaffolding generators.
Experience building or maintaining a frontend platform or design system — shared component library, design tokens, Storybook / visual-regression infra, a11y or perf budgets across apps.
Open-source contributions to Nx plugins, GitHub Actions, or related developer tooling.
Exposure to multi-tenant, regulated, or air-gapped deployment environments.

Responsibilities

Own the monorepo and build system
Own the architecture and ongoing health of our Nx-based monorepo: project graph, dependency boundaries, language toolchains (pnpm, uv, Go), and code generation patterns.
Make builds and tests fast, reliable, and incremental at scale.
Design and operate CI/CD
Build CI/CD systems with strong guarantees around correctness, reproducibility, and speed.
Own pipeline architecture, caching, parallelism, secret flow, and release engineering for environments that include regulated and air-gapped customer deployments.
Developer tooling and workflows
Design internal CLIs, scaffolding, local dev stacks (Tilt, Docker, Kubernetes), and code-quality infrastructure (lint, typecheck, format, pre-commit).
Own the developer-facing experience for deployment and environment workflows, while partnering with Infrastructure on the underlying cloud, Kubernetes, networking, and runtime substrate.
Create paved paths the team actually wants to use — not guardrails that get bypassed.
Treat developer tools as production systems
Run the build, CI, and dev infra with real SLOs, observability, and on-call rigor. Internal downtime is downtime.
Solve ambiguous productivity problems
Identify what's slowing engineers down, instrument it, and fix it from first principles.
Move from "X is slow" to a measured, deployed improvement with clear before/after numbers.
Operate this like a real product
End-to-end ownership: architecture, implementation, rollout, observability, on-call, and iteration based on internal user feedback.
Partner closely with every engineering team. Drive adoption through quality, not mandates.