Recall-Impact-Assessment-Agent

This is an LLM agent that identifies hospital inventories affected by an FDA product recall. The agent identifies the products at the hospital using a Cortex search and creates an executive summary to be shared with the hospital, along with the dollar value at risk if the product is not returned

🏥 FDA Recall Impact Assessment Agent

Built with Snowflake Cortex AI | Healthcare Supply Chain

The Problem

When the FDA issues a medical device or drug recall, the hospital supply chain teams manually cross-reference inventory, identify affected departments, and draft stakeholder alerts — a process that takes 2-3 days, and is prone to error.

The Solution

An AI agent that ingests an FDA recall notice, cross-references it against hospital inventory in real time, quantifies financial exposure, and drafts an internal alert — in under 60 seconds.

Architecture

FDA OpenFDA API → Snowflake Table → Cortex Search → Cortex Complete → Streamlit UI

Features

• Real-time FDA recall ingestion via OpenFDA API
• Natural language inventory matching using Cortex Search
• Financial exposure quantification by department
• Auto-generated stakeholder alert draft
• Interactive Streamlit dashboard

Tech Stack

• Snowflake Cortex (Search + Complete)
• Streamlit in Snowflake
• Python + Snowpark
• OpenFDA Public API

How to Run

1. Clone this repo
2. Set up Snowflake credentials in .env.
3. Run setup.sql in your Snowflake account
4. Deploy app.py as a Streamlit in Snowflake app

Data Sources

• FDA OpenFDA Device Recall API (public)
• Simulated hospital inventory (representative of real supply data)

Portfolio Context

Built as part of a Healthcare Supply Chain AI portfolio to demonstrate product thinking + technical execution.