SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 250,000 - 400,000 / annual
Legion Health is building autonomous medical care powered by AI, starting with psychiatry. The company has developed an AI-native care-delivery platform that automates 95% of administrative work and recently became the first company authorized to let AI prescribe psychiatric medications. Legion is backed by Y Combinator and has raised $23M+ with $5M+ ARR.
You will own entire problem domains end-to-end, not tickets or lanes within someone else's architecture. The role requires taking gnarly domains from "this is a mess" to "this is done and documented." You'll have written authority over domains selected together in week one, which may include:
- The factory spine: maximizing leverage in agentic engineering. Legion ships 500–1,000 PRs monthly with a lean team. You'll move the system toward 10,000 PRs/month while maintaining quality and rigor for a regulated clinical product that e-prescribes controlled substances.
- The clinical work spine: an event-sourced system of record for clinical work (orders, tasks, commands, receipts) that enables AI agents and humans to act on patient care with auditability and fail-closed checks.
- Agent operations in production: LLM agents performing real clinical-adjacent work with tool use, checks, and receipts. You'll decide how autonomy expands as proof accumulates.
- The data and metrics estate: warehouse, analytics platform, and pipelines that provide visibility into what actually happened.
- Architecture syncs with the CTO: you'll have standing to disagree, make calls in your domain, and unpack decisions for the team.
The company runs a regulated clinical product on an event-driven TypeScript platform and ships at unusual velocity with heavy agent leverage. Engineers who thrive here take ownership of complex domains, disappear into them, and return with working, documented solutions. The role is not for people who need well-scoped tickets, prefer management, or expect big-company pace.
Tech stack: TypeScript, Node.js, Postgres, Next.js on Vercel, AWS, production LLM agent orchestration, bespoke harnesses, tools, CLIs, skills, evals and observability, EHR and e-prescribing integrations, PHI-grade security and audit trails.
The interview process includes a 20-minute coffee chat, a 60-minute systems deep dive, an async take-home or 2–5 day paid work trial on a real domain from the backlog, interviews with founders and engineering, and a decision within 48 hours.
REQUIREMENTS:
- 10+ years of experience (though scars matter more than years)
- Owned architecture across multiple companies or many codebases: staff-plus IC, chief-architect, or platform-consulting background
- Ship daily at high volume with a fleet of AI agents (not just using Copilot occasionally)
- When handed an ambiguous, ugly, load-bearing problem, take it whole: read everything, decide, build, write the doc, tell the team what you decided
- Communicate proactively in writing
- Hold clinical-grade correctness without losing speed: auditability, HIPAA, fail-closed defaults, invariants over vibes
- Seen multiple systems owned end-to-end; this is a senior seat
- Not looking to manage instead of build
- Not someone who needs well-scoped tickets to be productive
- Not someone for whom big-company pace feels normal
