Senior Software Engineer

Marvin Behavioral Health is building a production healthcare application that helps people navigate mental health and wellbeing. With partners including employers and health plans, members get one place to understand coverage, find care, and get support, including an AI coach built into the product. This senior full-stack role is for engineers who want AI woven into both the product and their day-to-day workflow, not treated as a novelty.

Location: Los Angeles, CA (hybrid, 2 days per week in office, LA or Denver)
Compensation: USD 160,000 - 190,000 per year + equity
Experience: 5+ years

What you’ll do

• Own features end to end across a Next.js/React/TypeScript app, a Prisma/Postgres backend, and AWS infrastructure.
• Extend Marvin’s AI coach and deliver new LLM-powered member experiences, including retrieval over benefits data, structured extraction, guided flows, and agentic tool use against your own APIs.
• Build evaluation and observability for AI features so you can detect when performance degrades, not only when systems fail.
• Advance AI-assisted development practices with agent instructions, reusable skills, hooks, MCP integrations, and review practices that maintain testing and reliability as velocity increases.
• Handle sensitive health data responsibly, including PHI boundaries, redaction, audit trails, and vendor data agreements as part of the design.
• Review code critically, including code generated by humans or agents, and mentor teammates to raise the bar on testing and reliability.

What you bring

• 5+ years building production web applications with deep TypeScript, React, and Node experience.
• Next.js experience is a strong plus.
• Comfort across the stack: relational data modeling (using Prisma and Postgres), auth flows, caching, background jobs, and cloud deployment with AWS, Docker, and Terraform (or similar).
• A real testing habit across unit, component, and end-to-end layers.
• Experience with Vitest, Storybook, Cypress, and Playwright.
• Strong product judgment and the ability to collaborate directly with design, clinical, and partner-facing teammates.
• Daily expert use of agentic coding tools such as Claude Code, Cursor, or Codex.
• Hands-on shipping experience with LLM features: prompt design and versioning, structured outputs, tool/function calling, streaming UIs, retrieval-augmented generation, and multi-turn conversation state.
• Run evals and trace production LLM traffic (examples include LangSmith, Braintrust, Langfuse, or homegrown), and explain what changed when a model version updated.
• Understand the cost, latency, and quality trade-offs and make real decisions involving model selection, caching, prompt compression, and fallback strategies.
• Clear understanding of limits: when not to use an LLM, how to constrain it when it is appropriate, and how to prevent hallucinations from reaching members.

How Marvin works

• AI is part of the engineering culture, not a side experiment.
• The repository includes agent configuration.
• Shared review practices support code review and security checks, shifting effort toward judgment over typing.
• No unreviewed AI-generated code lands, and “the agent wrote it” is never treated as a bug explanation.

Nice to have

• Healthcare, benefits, or regulated-data experience (for example HIPAA or SOC 2).
• 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 projects.

Compensation and benefits

• Base salary: $160,000 - $190,000
• Equity
• Health, dental, and vision
• [401(k)]
• Unlimited PTO
• Hybrid: 2 days per week in office (LA or Denver)

Technologies you’ll work with

TypeScript, React, Node, Next.js, Prisma, Postgres, AWS, Docker, Terraform, Vitest, Storybook, Cypress, Playwright, Claude Code, Cursor, Codex, LangSmith, Braintrust, Langfuse, MCP, Retrieval-augmented generation, and LLM systems. PHI concepts are referenced as part of the role.

About the role

This is a senior full-stack engineering position where AI is a core part of how you build and where you build AI into the product. It is not a machine learning research role: you will not train models or tune GPUs.