Senior Software Engineer

Marvin Behavioral Health•Los Angeles, CA

13d•$160,000 - $190,000•Hybrid

About The Position

Marvin is hiring a senior full-stack engineer to build and maintain its production healthcare application. The role works across a modern TypeScript, Next.js, and React stack, a Prisma/Postgres backend, and AWS infrastructure. It is focused on shipping product rather than machine-learning research, with AI integrated both into development workflows and member-facing experiences. The engineer will extend Marvin’s AI coach and build LLM-powered experiences, including benefits-data retrieval, structured extraction, guided flows, and agentic tool use with internal APIs. The role also establishes evaluation and observability for AI features, advances AI-assisted development practices, and ensures sensitive health data is handled responsibly through appropriate PHI boundaries, redaction, audit trails, and vendor data agreements. The position calls for strong full-stack engineering, product judgment, and experience delivering LLM features in production. Engineers review human- and AI-written code critically, mentor teammates, and help maintain high standards for testing and reliability. The role is hybrid, with two days per week in the office in Los Angeles or Denver.

Requirements

At least 5 years building production web applications.
Deep experience with TypeScript, React, and Node; Next.js is a strong plus.
Comfort across the stack, including relational data modeling, authentication flows, caching, background jobs, and cloud deployment.
Experience with Prisma and Postgres, or relevant relational data modeling skills.
Experience with AWS, Docker, Terraform, or similar cloud deployment tools.
A strong testing practice at unit, component, and end-to-end levels; the team uses Vitest, Storybook, Cypress, and Playwright.
Strong product judgment and ability to work directly with design, clinical, and partner-facing teammates.
Daily expert use of agentic coding tools such as Claude Code, Cursor, or Codex, including writing agent instructions and reusable skills, delegating multi-step tasks, and carefully reviewing outputs.
Hands-on experience shipping LLM features, including prompt design and versioning, structured outputs, tool/function calling, streaming UIs, retrieval-augmented generation, and multi-turn conversation state.
Experience running evaluations and tracing production LLM traffic using tools such as LangSmith, Braintrust, Langfuse, or homegrown systems.
Ability to explain the effects of model-version changes in production.
Understanding of cost, latency, and quality trade-offs, with experience making decisions involving model selection, caching, prompt compression, and fallback strategies.
Ability to recognize when an LLM is inappropriate, constrain its use when appropriate, and prevent hallucinations from reaching members.
Review all AI-generated code before it is merged; unreviewed AI-generated code must not be shipped.

Nice To Haves

Healthcare, benefits, or other regulated-data experience, including HIPAA or SOC 2.
Experience building MCP servers or internal agent tooling for a team.
Experience with Clerk, Stripe, or Sentry.
Contributions to AI developer tooling, open source or internal.

Responsibilities

Own features end to end across the Next.js/React/TypeScript application, Prisma/Postgres backend, and AWS infrastructure.
Extend the AI coach and build LLM-powered member experiences, including benefits-data retrieval, structured extraction, guided flows, and agentic tool use against internal APIs.
Design evaluation and observability for AI features to detect quality degradation as well as system failures.
Advance AI-assisted development practices through agent instructions, reusable skills, hooks, MCP integrations, and review practices.
Handle sensitive health data responsibly, including PHI boundaries, redaction, audit trails, and vendor data agreements.
Review code written by people or AI agents critically.
Mentor teammates and raise standards for testing and reliability.