Fully Agentic Revenue Cycle Management Startup – Remote-first

We are a very early-stage, founder-led startup building the first truly agentic, fully automated Revenue Cycle Management platform for a high-margin U.S. clinical niche in the U.S.

What You’ll Build (First 6 Months)

• Multi-agent systems (LangGraph/CrewAI/Autogen) that handle end-to-end RCM workflows: eligibility verification → claim creation → scrubbing → submission → ERA posting → denial management → appeal letter generation
• Fine-tuning and prompt engineering of Llama-3.1-70B/Mixtral/Qwen-2.5-72B on real 837/835/ERA data (we will provide cleaned datasets)
• RAG pipelines over payer policies, CMS guidelines, and clinic-specific fee schedules
• Evaluation frameworks for claims accuracy, denial prediction, appeal win rate
• Integrations with clearinghouses (Change Healthcare, Availity, Waystar) via X12/REST APIs
• Safe, auditable, HIPAA-compliant deployment (we’ll use Fireworks.ai/Together.ai/Groq + E2B sandboxes initially)

Requirements (Must-Have)

• 4+ years of hands-on ML/AI engineering experience
• Proven experience shipping production LLM applications (agents, RAG, fine-tuning)
• Strong Python + LangChain/LlamaIndex + at least one agent framework (LangGraph preferred)
• Real-world experience with U.S. healthcare data — 837/835 files, UB-04/1500 forms, ERA 835 parsing, payer rejection codes, etc. (this is non-negotiable)
• Previously worked at a health-tech or RCM company (Waystar, AKASA, Thoughtful AI, Olive, R1, Experian Health, etc.) OR built revenue-cycle automation at a provider organization
• Comfortable owning the entire AI stack in a 2–4 person team

Send:

1. Your CV
2. 2–3 sentences on the hardest healthcare-AI problem you’ve solved
3. A link to code/agent demos you’re proud of (GitHub, private repo, video — anything)
4. Your notice period and salary/equity expectations

We move extremely fast — expect a technical interview within 48 hours of a strong application.

Looking forward to building the future of medical billing with you.