Software Engineer - AI Agents & Health Data

This is an early-stage, high-ownership engineering role at a mission-driven health tech startup — one of the first two engineers after the Head of Engineering. That's a compelling career inflection point: you'll shape the platform architecture, not inherit it.

Compensation context: $140k–$165k for 3–6 years of experience is competitive but not top-of-market for senior engineers at well-funded tech companies. However, the equity upside at an early-stage public benefit corporation could be meaningful if the company scales.

Skills you'll build here are in high demand:

• LLM agent engineering (MCP servers, Claude Agent SDK, tool-calling) — one of the hottest and most transferable skill sets in 2024–2025
• Big Query + dbt data modeling at production scale
• Full-stack ownership at a regulated startup — rare resume signal

Green flags: flat structure with explicit growth through scope (not politics), greenfield work, autonomy to scope and ship, and a mission that's easy to get behind. The "automate your own job" culture signals intellectual honesty and a growth mindset.

Watch-outs: "flat by design, no ladder to climb" may mean limited formal promotion paths. Early-stage startups carry inherent risk. The annual background and exclusion screening is standard for government-adjacent health work but worth noting.

Best suited for engineers who want broad ownership, AI/ML tooling exposure, and health tech domain expertise — a strong foundation for future senior or staff roles.

Aradigm is a public benefit corporation helping people with cancer and other serious diseases access life-changing cell and gene therapies that can potentially cure them. We do this through a new coverage model for self-insured employers and health plans, and behind that model is a lot of data: claims and eligibility feeds, drug pricing, epidemiology, and the reference data that ties it all together.

We’re hiring two Software Engineers, among the first engineering hires after our Head of Engineering, to build the platform that runs Aradigm. Our engineers aim to automate their own jobs away so they can move on to the next big engineering challenge. That takes many forms: implementing a new data feed, building an internal tool, or shipping an agent that runs a recurring workflow like an opportunity analysis end to end.

Most of the work is greenfield, on a nascent platform you’ll help shape.

You’ll own business outcomes, not just code, with the autonomy to match: you’ll scope problems with the people you serve, set priorities, and ship end to end. You’ll report to our Head of Engineering and work alongside our data team. We’re flat by design: no ladder to climb, and growth comes from scope and ownership. We work remotely within the US, with a strong preference for the East Coast and travel to our NYC office 3–5 times a year.

Within your first year, you’ve built at least one agentic automation that the team uses every day.

• Build data infrastructure: implement new data feeds, and model and serve claims, eligibility, pricing, epidemiology, and reference data in Big Query with pipelines that are tested, observable, and trusted
• Automate the work, including your own: when a task repeats, turn it into a tool or an agent workflow instead of doing it again
• Build agent harnesses: MCP servers, tools, context, evals, and guardrails that let existing models like Claude reliably handle recurring work. We don’t build or train models; we build what lets them do useful work safely
• Build internal tools: create the tools our clinical, actuarial, operations, sales, and finance teammates use daily, starting from their problem rather than a spec
• Ship with…