Challenge
Medical billing rules evolve over time and need to be translated into reliable system behavior. The project explored how an AI-assisted workflow could help clinics interpret new requirements and prepare rule changes.
Role
I worked as an AI engineer responsible for the solution design and core implementation.
Contributions
- Designed and built an AI assistant for medical billing and coding workflows.
- Created a transformation flow for converting requirements into rule-code structures.
- Built JavaScript generation capabilities for the existing billing system.
- Applied prompt engineering with OpenAI, LlamaIndex, and Python.
Approach
The implementation treated generated output as part of a controlled engineering workflow, with structured transformations between source requirements and executable rules.