About Varick. Varick builds AI agents that take over real operational workflows inside the world's largest enterprises. Our forward-deployed engineers and strategists embed with client teams, map how work actually runs, then our engineers build the agents into production. We're venture-backed, revenue-generating from day one, and already in production inside several of these companies. You'll join an elite team from Meta, AWS Bedrock, Citadel Securities, McKinsey, BCG, Stanford, and more.
The Role. You'll build Varick OS and the agents that run on it: the runtime that executes long-running workflows against enterprise ERP and CRM environments, the retrieval and context layer those agents reason over, the eval and trace infrastructure that keeps quality measurable, the model layer that keeps us provider-agnostic, and the deployment tooling that ships all of it into environments we don't control. The surface area is large, and decisions you make in your first quarter will still be load-bearing in two years. You'll work alongside our lead engineer, who built the current stack from scratch. This is not a client-facing role: our forward-deployed team runs the audits and interviews, then hands you the workflow spec. You build it, and pull whatever repeats back into the platform so the next client is faster.
How We Build. We run an agent-driven development workflow: thorough specs up front, PRs graded against those specs, and eval suites that gate what ships. It's how a small team holds velocity without the quality drift that usually comes with it.
What You'll Do.
- Build the agent runtime. Long-running workflows against enterprise systems that go down mid-run: durable execution, safe retries, idempotency, retrieval and context assembly, and human approval gates that hold up when a step touches a client's general ledger.
- Own the model layer. Provider-agnostic routing with real fallback, cost and latency budgets, and no lock-in to any single vendor.
- Build evals and tracing. Golden sets, regression suites, and traces over every intermediate step, so we catch a drop in agent quality before a client emails us.
- Ship into client environments. Package, deploy, version, and upgrade the platform across single-tenant deployments, including into a client's own cloud account.
- Turn audits into production agents. Take workflow specs from our forward-deployed team, interrogate them, build them, and generalize the parts that repeat.
What We're Looking For.
- 3+ years building production software, with real ownership of systems other people depended on
- Strong backend and distributed-systems fundamentals: state, concurrency, idempotency, failure handling
- Experience operating production systems at scale: Kubernetes, cloud infrastructure, release management across versions
- At least one LLM system shipped to real users, and a clear account of how it failed in production
- A working understanding of inference economics, evaluation, and model behavior
- A habit of interrogating a spec instead of implementing it obediently, and comfort with fast, low-process development
- A track record of digging into code, data, and root causes directly, and strong communication with technical and non-technical people alike
Nice to Have.
- Shipping software into environments you don't control: BYOC, self-hosted, on-prem, or air-gapped
- Durable execution or workflow orchestration (Temporal or similar)
- Enterprise environments with strict security and reliability requirements (SOC 2)
- Deep familiarity with an enterprise ERP or CRM system of record: SAP, Oracle, NetSuite, Salesforce, Workday
- Founding or early engineer experience at a startup
What This Role Isn't.
- Not a research role. You won't train or fine-tune models.
- Not prompt engineering. Prompts are a small fraction of what makes an agent work in production.
- Not client-facing. You won't run audits, sit in discovery interviews, or manage stakeholders.
What We Offer. Meaningful equity (0.1-0.5%) at a stage where it still compounds, real ownership of the architecture decisions still open, scope that grows with every deployment, a team that hires up, flexible PTO, free lunch and dinner in the office, Ubers home if you're staying late, and monthly team dinners.
Logistics. On-site in our San Francisco Financial District office, 5-6 days a week. Comfortable with startup hours, for startup upside.