Founding AI Engineer

About LinkMD

LinkMD is building the AI operating system for surgical organizations — ASCs, office-based labs, and surgical groups. Our platform is a unified suite of specialized AI agents that run the full revenue cycle: eligibility, prior authorization, CPT coding, operative note generation, claims scrubbing and submission, payment posting, denial management, and patient collections.

We have live clients generating real revenue and real volume.We're early-stage, well-positioned, and growing fast. The US healthcare revenue cycle is broken in ways that cost providers and patients real money every day. We're fixing it at the infrastructure layer.

What You'll Work On

You'll own the AI layer of LinkMD end-to-end. Working directly with the founder on the hardest technical problems in the company:

* Architect and ship the reasoning brain that powers our platform — a hybrid system combining structured knowledge stores, graph databases, vector retrieval, and agentic orchestration

* Build and improve specialized AI agents across the revenue cycle (Prior Auth, CPT coding, Op Note generation, Claims, Denial Management, Collections)

* Design and maintain the ingestion pipelines that keep our knowledge base current — regulatory rules, payer policies, coding guidelines, client historical data

* Build evaluation infrastructure that lets us measure whether the system is actually getting smarter over time

* Make architectural decisions that define LinkMD's long-term technical moat

This is not a prompt-engineering role. It's not an LLM-wrapper role. You'll own how our AI actually thinks.

Qualifications

* Strong engineer with meaningful experience shipping LLM-based systems in production

* Deep fluency in Python and modern AI tooling

* Comfortable going from architectural whiteboard to production code without hand-holding

* Strong intuition for retrieval systems, agentic frameworks, and evaluation — you understand where naive approaches break

* Track record of owning hard technical problems end-to-end

* Bias toward shipping fast, learning faster, and taking full ownership of outcomes

* Energized by complex domains — you see regulatory complexity as an interesting problem, not a blocker

Nice to Have

* Healthcare, RCM, insurance, or regulated-domain experience

* Experience with knowledge graphs, hybrid retrieval architectures, or multi-agent systems

* Exposure to CMS data, payer policies, X12 EDI, FHIR, or HL7

* Prior startup experience, especially at the founding or early stages

* Fine-tuning experience with sound judgment about when retrieval beats fine-tuning

We care more about raw ability and judgment than specific framework experience. If you've built hard things, you'll pick up the rest.