Staff Software Engineer (Payer)

About this role

• Payers & Outcomes at Headway sits at the intersection of ML, clinical quality, and trust
• You’re building the intelligent systems that protect the integrity of our network — spotting risk, surfacing quality signals, and giving providers real-time feedback so they can deliver great care. What you could work on:
• Clinical Feedback: Architect the LLM-powered systems that review clinical documentation at scale and deliver helpful, real-time feedback to providers. Turn a traditionally painful audit process into a coaching experience — and scale it across the entire network
• Detecting Risk in Healthcare Claims: Design the ML/LLM systems that spot patterns of inappropriate billing, over-utilization, and low-quality care across millions of appointments and claims. Translate those signals into clear feedback for providers and transparent reporting for our insurance partners
• Measuring & Elevating Clinical Quality: Help build the systems that systematically review every prescriber in our network against clinical quality standards — so we can measure, improve, and defend the quality of care Headway delivers. The stakes are real: our most recent launch reviews clinical notes and nudges providers to put a safety plan in place when a patient shows signs of suicidal risk — a system projected to save roughly 1,000 lives a year
• Patient Needs & Provider Fit: Crack the “one size fits all” problem. Build ML systems that understand how complex a patient’s needs are and match them to providers best suited to help — so every patient gets the right care and every provider operates at their best
• Trustworthy AI in Healthcare: Champion how Headway evaluates and iterates on AI systems. Build the evaluation frameworks, optimization loops, and observability that keep LLM-powered features reliable and safe in a regulated environment
• Technical Foundations & Mentorship: Set the technical bar for the Payers & Outcomes group — ~12 engineers today, growing to 30–50 by end of year. Lay the ...