- Client
Halo ↗
- Technologies
TypeScript, Python, Postgres (Supabase)
Overview
Rebate and pricing engine for a pharmacy client. It reads the daily wholesaler files, matches each drug to its contract, and shows every pharmacy where it stands on its rebate tier and what it takes to reach the next one. Rules do the math, an LLM only explains it. The engine recommends, a person decides.
Technologies
- TypeScript
- Python
- Postgres (Supabase)
- LLM explanations
- Rules engine
Key Features
- Reads the daily wholesaler files and matches each drug to its contract
- Shows each pharmacy its rebate tier and what it takes to reach the next one
- Deterministic rules compute every number, the LLM only explains them
- Every output cites its source rows and the rule version
- Per-client data isolation and full traceability
Challenges & Solutions
Math you can audit
Rebate money cannot depend on a model's guess. Rules compute tier compliance and net price, and the LLM only explains the result in plain words. Every output points back to its source rows and rule version.
An engine that recommends, never executes
The engine shows where a pharmacy stands and what to do next. It never acts on its own. A person makes the call.
Internal tool, public client site
The engine runs inside the client's operations, so this page shows their public site, not the app. No client data appears here.
My Role
Software engineer on the AI product team at Resorsi. Built the file ingestion, the drug-to-contract matching, the tier compliance and net price logic, and the recommend-only boundary.
Results
The engine shows each pharmacy its rebate tier and the gap to the next one, with every number traced to its source. Built at Resorsi. It runs in a staging environment, not in production. No public app URL.