Senior Platform Engineer

Sprig•San Francisco, CA

2d•$180,000 - $260,000•Hybrid

About The Position

Sprig is seeking a Senior Platform Engineer to own the build-and-test loop for their AI-powered customer survey platform. This role is crucial for ensuring that development changes are quickly and reliably validated, impacting the speed and trust of the entire organization. The engineer will be responsible for modernizing the build systems, reducing CI time and cost, improving observability, enhancing test automation, transitioning to ephemeral environments, and building an agent-ready development platform. The Platform team manages systems supporting over a billion visitors per month with traffic peaking at 150k requests per second. Sprig is a fast-growing company backed by prominent investors, focused on making deep customer understanding effortless and always on.

Requirements

6+ years owning a monorepo build system at scale (Bazel strongly preferred; Buck2, Pants, Nx, or Turborepo also considered).
Understanding of hermeticity, sandboxing, action graphs, and cache invalidation from first principles.
Experience owning incremental builds, test sharding, self-hosted runner fleets, presubmit/postsubmit design, and merge queues.
Deep experience with production bundler pipelines (webpack, Vite, esbuild, or Turbopack), shared UI packages, and test suites (Jest, Playwright, Cypress) and their interaction with hermetic builds.
Strong expertise in Node.js module resolution and packaging (CJS/ESM, exports).
Comfortable operating in a polyglot repository alongside Go.
Experience eliminating hidden network calls, machine-dependent toolchains, and undeclared inputs to ensure reproducible results and prevent cache/sandbox poisoning.
Proven ownership of flake detection/quarantine, balanced sharding, and merge signals, with measurable results in build times, CI spend, and artifact or bundle sizes.

Responsibilities

Modernize the JavaScript and TypeScript build, including migrating from current Node Bazel rules to a maintained successor, moving the JS toolchain to Bazel modules, and replacing workarounds with proper module linking and incremental compilation.
Cut CI time and cost by establishing and driving down p50 and p95 PR build and test times through improved cache hit rates, incremental builds, and balanced test parallelism. Own the remote cache fleet's sizing, eviction behavior, and cost.
Make the build observable by shipping telemetry for per-target timing, cache hit rate, flake rate, and cost to enable data-driven build decisions.
Own the developer experience for test automation across unit, integration, browser end-to-end, and visual layers, ensuring local commands match CI and making it easy to run specific tests. Detect and quarantine flakes.
Replace shared, contended environments with short-lived, isolated stacks that can be created per branch or task by humans or cloud agents, focusing on fast spin-up, real isolation, and cost control.
Build the agent-ready development platform, delivering hermetic, sandboxed, reproducible build and test environments for agents, along with automated review and presubmit signals.