Platform Engineer

About the role Cityblock is betting that agents, operating our systems safely and under real oversight, can meaningfully expand what our engineering and care teams can accomplish using AI. This role builds the platform that makes that possible. You’ll create and mature the platform that every other engineer builds their agents on. This is a high-ownership, high-leverage role that shapes how the whole company scales. You'll report to the Director of Engineering, Platform, working alongside, learning from, and up-skilling platform team peers. This is an AI and agent platform role first. You design the harness, sandboxing, permissions model, and observability that let agents act on internal systems safely. You will decide what an agent can touch and under what supervision; and you give every automated action an audit trail. When another team ships an agent, the foundation they rely on is the one you built. Because that platform runs on cloud infrastructure, you also bring a deep working understanding of it: GCP, Terraform, and the reliability, security, and compliance practices that keep a regulated health platform running. You will Build the AI foundation Build and maintain the agent harness other teams use: sandboxing, RBAC, audit trails, and the runtime agents act through. Make internal systems API-ready for agents, and design the MCP servers, context systems, and agent registry that expose them safely. Define autonomy tiers: what an agent can do unsupervised, what requires human review, and how an action gets rolled back when it shouldn't have happened. Keep the infrastructure sound Shape and hold the bar for the reliability, security, and compliance posture of the platform your work runs on, partnering with the rest of the platform team on day-to-day operations and shaping the GCP and Terraform infrastructure where the agent platform requires it. Automate operational toil so the platform stays fast to build on as it scales. Work across teams Work as a peer to the platform team and a partner to every engineering team that builds on your foundation. Partner with security to make agent autonomy and audit defensible in a regulated environment. Work with the data platform so agents act on trustworthy, well-governed data. Requirements for the role 8+ years of experience as an engineer You are a hands-on builder of agent systems, with the infrastructure depth to run them well. Hands-on experience building agent orchestration or tooling including MCP, agent runtimes, harness design, or comparable systems. Deep working knowledge of cloud infrastructure and infrastructure-as-code (GCP and Terraform or equivalents). You have designed for safety and governance: RBAC, sandboxing, audit trails, autonomy boundaries, and rollback. You've built in a regulated or high-availability environment where safety and uptime are non-negotiable. HIPAA specifically is a plus, not a requirement. Security or infosec depth beyond standard platform hygiene. Senior Staff-level judgment: you set the architecture others depend on and own a platform end to end. Rock-solid engineering fundamentals and architectural instinct that stand on their own, independent of AI: you reason about distributed systems, failure modes, data flow, and design tradeoffs from first principles. Fluent daily use of AI assistants for engineering, with a clear sense of when to trust them and when not to. Comfortable with ambiguity and defining your own scope. Part of the job is deciding what the platform should become. How we define success Engineering teams build and ship agents on your platform, and adoption grows without one-off help from you. Agent actions across the stack are safe by default: scoped permissions, clear autonomy tiers, and a complete audit trail for every automated action. Infrastructure stays reliable, secure, and compliant as the platform scales. Nice to have, but not required A CS degree or equivalent. Direct HIPAA or SOC 2 experience We