Firmwork is building the execution layer for AI-native firms.
Expert firms still run on inboxes, folders, documents, spreadsheets, meetings and people's memory. AI has entered the firm, but mostly as a tool sitting outside the real work: a chatbot, a drafting assistant, a search box.
That is not enough.
The real shift is turning the firm itself into an AI-native operating system: matters, documents, decisions, approvals, workflows and institutional knowledge connected in one controlled layer where agents and humans can work together.
Firmwork is building that layer.
We started in legal, where the pain is obvious: M&A deals involve thousands of documents, complex diligence, drafting, negotiation, status tracking, client updates, approvals and high-stakes judgment. Our platform gives teams a workspace where agents can read the matter, operate on documents, produce work product, cite sources, propose changes, branch/merge work, and stay under human supervision.
Now we are deploying this inside real firms and adjacent expert workflows: law firms, private capital, family offices and operating businesses that want to become AI-native from the inside.
We are hiring a Founding Engineer to build the infrastructure behind that.
What you'll build
You will work directly with the founder to build agentic infrastructure for real professional work.
This means:
- Agent runtimes that can operate over real files, workflows and user intent.
- Virtual file systems for matters, projects and firm knowledge.
- Branching, merging, review and approval flows for AI-generated work.
- Durable workflows with retries, persistence, observability and human checkpoints.
- Sandboxed execution environments for agents.
- Secure multi-tenant infrastructure with permissions, audit logs, secrets and data boundaries.
- Document pipelines for PDFs, DOCX, spreadsheets, data rooms and generated work product.
- Integrations with Google Drive, Microsoft 365, storage systems and internal firm workflows.
- Backend systems that make AI agents reliable enough to use in serious environments.
This is not a normal full-stack role.
The frontend and product surface matter, and we care deeply about taste. But the biggest need is infrastructure: backend architecture, security, reliability, workflow orchestration, agent execution, data models, permissions, deployment and systems thinking.
Your job is to double our technical velocity and raise the quality of the infrastructure underneath everything we ship.
What we are looking for
You might be a fit if you:
- Are excellent at TypeScript, backend systems and product engineering.
- Have built real software that had to be reliable, not just demos.
- Understand databases, queues, storage, auth, permissions, APIs, observability and deployment.
- Have strong instincts around security, tenant isolation, auditability and data boundaries.
- Are excited by AI agents, tool use, workflow engines, sandboxes and human-in-the-loop systems.
- Use AI tools aggressively in your own work, but do not confuse speed with sloppiness.
- Can sit with users, understand messy workflows, and turn them into clean infrastructure.
- Want founder-level ownership, not a narrow engineering ticket queue.
You do not need to know legal. You do need to care about understanding complex work deeply enough to automate it responsibly.
Our stack
We currently work with:
Next.js, React, TypeScript, Vercel AI SDK, Vercel Workflows, Vercel Sandbox, Supabase/Postgres, WorkOS, Pinecone, Upstash, Trigger.dev, Google/Microsoft APIs, PDF/DOCX tooling and custom agent infrastructure.
You do not need to have used every tool. You need to be able to reason from first principles and learn fast.
First 90 days
Days 1-30: Understand the Firmwork framework, the product, current deployments, user workflows, the agent runtime and the infrastructure underneath it all.
Days 31-60: Own a specific slice of production, taking one runtime, workflow or product surface from problem framing through implementation, deployment and review.
Days 61-90: Spearhead real innovation in the legal field around agentic harnesses: own infrastructure deployment and the application layer end to end while shipping valuable production workflows for users.
How to Apply
To apply, please email santiago@firmwork.ai with your resume/GitHub/LinkedIn and a few sentences about why this work has your attention. We love shipped systems, deep technical notes, and unconventional projects.