The company builds an agentic AI orchestration and agent development platform. Customers use it to build, deploy and run AI agents and workflows in production, on a runtime that executes at the edge and scales on demand.
It is a small team of senior engineers, and each engineer shapes the platform directly.
Responsibilities
Own the platform every flow runs on, including the edge runtime, container execution, the data layer and the infrastructure beneath them.
Keep production fast and reliable, and lead on-call.
Turn a fast-growing system into one that is reproducible, observable and safe for agent-driven development.
Work directly with the CTO on architecture decisions, with a clear route to leading platform engineering as the team grows.
Own the edge runtime: Cloudflare Workers, Durable Objects, Queues and dead-letter queues, KV and D1, the dispatcher and the deploy pipeline.
Own container execution, covering how flows run, scaling, retries and capacity.
Own production reliability: on-call, incident response, alerting, dashboards and postmortems.
Own security: credential encryption, secrets management and compliance work.
Own usage metering, so credit-based pricing is tracked and reported accurately.
Design, build and run the services that execute customer agents and workflows at scale.
Lead incident response, find root causes, and make sure each incident leaves the system stronger.
Move infrastructure into code across Cloudflare, Vercel, Supabase and DigitalOcean.
Build the testing, sandbox and observability foundations for agent-driven development.
Review designs and code, mentor engineers, and raise the bar for reliability.
Within 30 days: understand the system end to end, deploy, roll back and resolve a production incident on your own, and join the on-call rotation.
Within 60 days: have alerts cover the parts of the platform that actually fail, give every critical service a runbook, and ship your first infrastructure-as-code pilot.
Within 90 days: have credit-based usage metering live and accurate, and put sandbox environments and quality gates in place so engineers and coding agents can ship changes safely.
Requirements
5+ years building and running production backend systems, with real on-call ownership.
Strong TypeScript and Node.js.
Agent-driven development (ADLC): daily work with Claude Code or similar coding agents, running loops that plan, implement, test and verify changes, and writing the instructions, guardrails and checks that keep them reliable.
Production experience with Cloudflare Workers, including Durable Objects, Queues, KV and D1, and their limits and failure modes.
Working knowledge of Alchemy for TypeScript infrastructure-as-code, Effect TS for typed errors, dependencies and tracing, and e2e.com or a comparable end-to-end testing tool. Show a project you built with at least one of them.
Experience with containers, queues and distributed systems, and the failure modes that come with them.
Comfort with Postgres. Experience with an analytics store such as ClickHouse is a plus.
A habit of writing runbooks, design notes and postmortems.
Other infrastructure-as-code tools, such as Terraform or Pulumi.
Observability with Grafana, OpenTelemetry or similar.
Vector databases such as Weaviate or Qdrant, LLM gateways such as LiteLLM, or data integration tools such as Airbyte.
Usage-based billing with Stripe.
Security or compliance work, such as ISO 27001.
Conditions
The team is moving to an agent-driven development lifecycle in which engineers direct coding agents and every change is tested and verified in an isolated sandbox before review. You will help build that workflow, not only use it.