Founding Engineer

Description

The company is building an AI-native immigration law firm for highly skilled workers and the companies that employ them.

Employment-based visa cases such as O-1s, EB-1s, L-1s, H-1Bs and similar petitions are still handled through slow, costly, manual steps. The company uses AI for document analysis, drafting and workflow automation, with licensed attorneys reviewing the work, to prepare USCIS-ready petitions much faster.

Matthew, the CEO, previously founded a legal software company that was named to Forbes 30 Under 30. He and Rachel, the COO, are siblings who grew up around immigration law through their father's practice of more than 40 years. Claire, the CTO, has built AI and compliance systems at scale.

This is a Founding Engineer role working across client experiences, internal tools, backend services, and AI agents and workflows.

Responsibilities

• Build and improve the stack that runs the product: Next.js and TypeScript on the frontend and app layer, Supabase and Postgres for data, AI workflows, document generation, and internal review tools.
• Ship everyday full-stack work: features, bug fixes, and improvements to existing flows, taken from a vague request through design, build, testing, and release.
• Build intake, evidence collection, case dashboards, document uploads, and notifications for clients and internal staff.
• Investigate bugs and fix reliability and performance problems across the app, and improve usability.
• Talk with attorneys and operations to learn how cases really move, then turn manual steps into practical software.
• Build agent capabilities: tool use, retrieval, orchestration, state management, and the code that keeps agents working in production.
• Build the workspaces and sandboxes agents use to read, create, and organize files, keep their output, and recover from failures.
• Improve AI-assisted document work, including processing, evidence mapping, case drafting, follow-up questions, and generation.
• Make AI output easier to check: source attribution, version history, review screens, and human checkpoints for attorneys.
• Work with the legal team on evaluations, use failures to guide fixes, and add tests and observability to what you ship.
• Propose approaches, explain tradeoffs, and agree on scope and technical decisions with Claire and Matthew.
• First 30 days: learn the product, codebase, case workflows, and attorney review process. Ship useful fixes and features, work across the whole stack, and find where clients and the team struggle.
• By 60 days: take one meaningful workflow from problem to release, work directly with operations or attorneys, make sound tradeoffs, and act on feedback after launch.
• By 90 days: deliver consistently across product and AI work with less supervision, with measurable gains in manual effort, reliability, review quality, or client experience.

Requirements

• 3+ years building production web applications, or equivalent proof of strong product engineering.
• Strong TypeScript, React, and full-stack skills. You can work across frontend, backend, APIs, databases, and internal tools.
• Good frontend judgment. You care about clear interactions and make complex workflows easy to use.
• Sound systems judgment. You can reason about data models, permissions, integrations, background jobs, and failure modes without overengineering.
• Independence. You can turn a vague request into a plan, ship it, and own how it behaves in real use.
• Clear communication. You ask good questions, explain tradeoffs, take feedback, and raise problems early.
• A strong drive to learn agent systems and the immigration workflows the software supports.
• Strongly preferred: experience building agent systems. It is not a hard requirement. Strong full-stack engineers who show solid fundamentals, curiosity, and fast learning are encouraged to apply.
• Especially useful experience includes agents that use tools, retrieve information through RAG, manage state, and run multi-step workflows.
• Agent workspaces, filesystems, sandboxes, and safe execution of generated code or tool calls.
• Production AI features such as structured outputs, streaming, retries, observability, evaluations, and human review.
• Frameworks or protocols such as Vercel AI SDK, Mastra, MCP, or comparable tools.
• Tell us what you have built, the decisions you made, and what you learned when something failed.

Conditions

Full-time, based in San Francisco.

Visa sponsorship is available.

Benefits

• Optional extras, if you have experience in these areas: work on document-heavy products or complex operational workflows; work in legal tech, fintech, healthcare, HR tech, compliance, or another field where correctness and sensitive data matter; early-stage startup experience, substantial side projects, or meaningful open-source contributions.
• Ownership of core product systems in a small, early team with direct access to users, attorneys, and operators.