Software Engineer-Patient Products

About the role

At Legion Health, we are developing the AI infrastructure to deliver fast, affordable, and world-class health care at world scale. The role owns the core product and agent systems that coordinate patients, clinicians, and AI agents as one coherent machine. You will architect and scale the event-driven backend while shaping the workflows and guardrails that make autonomous psychiatric care reliable and usable. This position requires deep ownership of end-to-end product behavior, from state modeling and safety rails to integrations and human + AI UX. You are expected to evolve with the system, carrying real weight and growing into running major parts of the production environment.

Key facts

What you'll do

• Own our event-driven backend - Architect and scale our Node.js / TypeScript / Supabase (Postgres) / AWS stack, designing schemas, invariants, and workflows that encode how psychiatric care actually operates and turn messy real-world processes into clean state machines and event streams.
• Build real LLM agents as coworkers - Implement tool use, retries, memory, and safety rails, designing action schemas and evaluation loops so agents can run reliably in production and orchestrate context and multi-step workflows effectively.
• Shape human + AI ops UX - Build internal tools used by both humans and agents, making it trivial to see "what happened, why, and what should happen next" in any patient journey while keeping the experience coherent and actionable.
• Define world-state & simulation - Model the canonical state of a patient's journey across time, powering alerting, routing, and decision-making from that live simulation to reflect clinical reality accurately.
• Own data, safety & compliance - Engineer HIPAA-compliant pipelines for transcripts, events, and EHR data, ensuring PHI access, agent actions, and human overrides are all auditable and meet regulatory expectations.
• Drive architecture & strategy - Work directly with leadership to debate tradeoffs, define new primitives, and decide what we build next in a system that must scale safely and remain understandable.
• Collaborate closely with clinicians - Translate clinical processes into executable workflows and state machines, ensuring that the software faithfully supports complex psychiatric care operations.
• Iterate rapidly on reliability and correctness - Move quickly while maintaining robustness, using tight feedback loops and high standards to keep the system trustworthy and performant.

Requirements

• You are authorized to work in the United States without sponsorship now or ever, and you are not requiring or expecting work authorization sponsorship or changes in employment for this role.
• You are located in the San Francisco Bay Area or are willing to relocate there immediately and maintain presence there for the duration of the role.
• You have the right to permanent employment in the United States and are not requiring or expecting work authorization sponsorship or changes in employment for this role.
• You are not on a visa that would be impacted by working in California or by the nature of this role, and you confirm you are not requiring any changes in your current work authorization.
• You are not a student, intern, or trainee, and you have the ability to work full-time in San Francisco on an ongoing basis.
• You have direct, hands-on experience building and operating backend systems that are event-driven, stateful, and scalable in production environments.
• You are either already LLM-fluent or a strong systems/backend engineer who can become deeply effective with large language models and agent tooling very quickly.
• You have owned real systems from 0→1 or 1→N, demonstrating the ability to handle end-to-end responsibilities without needing a perfectly scoped box to operate inside.

Practical notes

Hours: Full-time (40 hours per week).

Visa: Not applicable per SOURCE requirements.

Travel: Not mentioned in SOURCE.

Deadline: Not mentioned in SOURCE.