Pfizer Externship via Extern·🟢 Open to workPortfolio and notes from my projects, process, and interests.
A curated set of projects and pieces I have completed, with context on what I built and why.
Collaborated on an interactive dashboard comparing eight Illinois agencies across stop volume, search rates, race- based search rates, average stop duration, and time trends. Implemented multi-agency selection and linked highlighting across four coordinated views, enabling users to explore agency comparisons through metric toggles, tooltips, and selected-agency summaries.
.jpg)
Pfizer · ✅ Verified by Extern · ⏱️ In progress
Prototype AI-powered document intelligence with Pfizer—using OCR, LLMs, and RAG to automate real enterprise PDF workflows and build a standout portfolio project.
Portfolio and notes from my projects, process, and interests.
My name is Paul Umanzor. I am a creative professional sharing my work and background here.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
In progress
Experience
Sales Support
Fastenal · Aug 2025 - Present
Information Technology Intern
SupplyCore · Feb 2023 - Sep 2023
Education
Northern Illinois University | DeKalb, IL
Bachelor of Science in Computer Science · Class of 2026
Skills
Collaborated on an interactive dashboard comparing eight Illinois agencies across stop volume, search rates, race- based search rates, average stop duration, and time trends. Implemented multi-agency selection and linked highlighting across four coordinated views, enabling users to explore agency comparisons through metric toggles, tooltips, and selected-agency summaries.
✅ Verified by Extern · ⏱️ In progress
Prototype AI-powered document intelligence with Pfizer—using OCR, LLMs, and RAG to automate real enterprise PDF workflows and build a standout portfolio project.
The project developed a prototype AI pipeline that combined OCR, LLM processing, and retrieval-augmented generation to extract and interpret information from enterprise PDFs. The work mapped model capabilities and integration steps, and documented research findings about how LLMs process and generate responses from extracted text.
I produced a concise technical summary explaining LLM training on large text corpora, the Transformer attention mechanism, and how models are fine-tuned with human feedback for improved generations.
I summarized LLMs by describing their training on large text corpora, the Transformer mechanism that assigns attention across sequences, and how models are fine-tuned with human feedback to improve generations.
