Extern Pfizer

Coming Soonfullstack

Write Python scripts that pull text from messy PDFs, build and evaluate OCR pipelines that can read labels no human wants to touch, and create AI-powered search tools that answer questions across thousands of pages instantly.

🌐 Live Site 📁 GitHub RepoBy Kyle Parsotan · 2026

Problem

Deliver a working document intelligence system for real-world clinical supply workflows at Pfizer.

Solution

Deliver a working document intelligence system for real-world clinical supply workflows at Pfizer

Tech Stack

Frontend

NextJSShadnuiMaterialUIResend for email serviceUser dashboard for uploading the document or using the camera for uploading the medincine

Backend

Node.jsExpress.jsVector Database (chroma)Document Parsing: PyMuPDF + Tesseract OCR fallback

Database

PostgreSQLAWS S3 for cloud storageLocal storage

DevOps

CI/CDDocker containerization

Architecture

Future Improvements

→Stay up to date
→Work on optimizing and the building the full stack application
→Integrate this within the mediconnect fullstack application
→Share the results to other professional