All work
Healthcare AI

ChartGuard

A clinical AI platform that audits every documented visit before it bills, catching compliance risks and under-coded revenue that manual sampling misses.

Role
Full-stack development, clinical rules engine design, coding pipeline
Year
Selected work
Stack
React · Python (FastAPI) · PostgreSQL · pgvector · OpenAI · AWS

The problem

Behavioral health providers could only spot-check a fraction of their charts, leaving compliance gaps, under-coded visits and safety-relevant details unexamined.

The approach

ChartGuard reviews clinical documentation against payor policy, regulatory requirements and the organisation's quality standards before a claim goes out. Every flag and coding suggestion includes a traceable reason and supporting documentation.

  • Clinical note ingestion and structuring
  • Payor and regulatory rules engine
  • Under- and over-coding detection
  • Compliance quality scorecards

The model finds the gap. The clinician still closes it.

Outcome

Quality management moves from reactive sampling to proactive review before billing.

Built with

  • React
  • Python (FastAPI)
  • PostgreSQL
  • pgvector
  • OpenAI
  • AWS