Infrastructure Engineer

Flint•San Francisco, CA

25d•Onsite

About The Position

Flint is the agentic platform that ships pages from your brand and ad campaigns. AI doesn't replace marketers. It removes the work that should never have been theirs: building a one-off landing page for every keyword, every campaign change, every A/B test, then waiting on design and engineering to do it. Today we deliver on-brand pages on your own domain in minutes, with no rebuilds and no engineers. We're backed by Accel, Sheryl Sandberg (former Meta COO)'s venture fund SBVP, and Neo. The vision Sandberg instantly understood Flint's vision. "I was sharing how it took (me) five teams three months to build one A/B test just to increase conversion by 10% on our Google ad," Lim said. "And then she stopped me, [and] said, 'Michelle, it was 140 people at Meta who had to do this'." — TechCrunch Imagine a website that builds new pages, tests creative, and optimizes performance in response to real-world signals: competitive shifts, user behavior, trending search queries, the way an autonomous vehicle responds to road conditions and traffic. The marketer still owns the work that compounds and the work that requires taste: strategy, audience, creative direction, brand judgment. Flint takes the manual production off their plate so they get those hours back. The role Flint runs background agents that make changes to real codebases and serves AI-generated pages at scale. That creates an infra problem most companies don't have: many agents editing code in parallel, and you have to coordinate, merge, and ship their work without it turning into chaos. You'll own the systems that power all of it: cloud architecture, CI/CD, deployments, observability, and security. It's a 0-to-1 role: you'll inherit a scrappy setup and make it production-grade. Recent work in this role: synchronizing and merging the git output of parallel background agents, building observability for agentic workloads, and hardening deployment pipelines.

Requirements

Scaled a product through real user growth and know what breaks, and when.
Strong with cloud (AWS or GCP), containers (Docker/Kubernetes), and IaC (Terraform or Pulumi).
Think in systems, performance, reliability, and cost trade-offs come naturally.
Can drop into infra you didn't build and make it better without a runbook.

Nice To Haves

Been through a startup growth phase.
Security depth: SOC 2, threat modeling, or pen testing.
Built self-service tooling that product engineers actually used.

Responsibilities

Own the concurrency and coordination problems of running many background agents in parallel, the core hard problem of this role.
Own cloud infrastructure end-to-end: architecture, provisioning, CI/CD, deployments.
Cut latency and cost as we scale, without trading off reliability.
Own observability and incident response, so we know when things break before users do.
Own security across the stack and deployment pipelines.
Make it easy for product engineers to deploy confidently and debug fast.