Currently an Extern @Pfizer
First-year at WashU St. Louis, creating AI prototypes for real document workflows and growing a professional network.
Portfolio items include an externship prototype that used OCR, large language models, and retrieval-augmented generation to extract data and build document insights for enterprise PDFs.
.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.
First-year at WashU St. Louis, creating AI prototypes for real document workflows and growing a professional network.
I am a first-year Biology student at Washington University in St. Louis exploring tech and AI through hands-on projects. I have completed an externship prototyping AI document intelligence, and I am focused on building relationships and expanding my professional network as I shape my early career.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
Pfizer
Experience
Pediatric Internship Program (PIPS) Research
Stanford University - Center for Academic Medicine · June - August 2025
Geoscience Specimen Collection Intern
Stanford University - Mitchell Earth Sciences · June - August 2024
Skills
✅ 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 work built a prototype document-intelligence pipeline combining OCR, computer vision, large language models, and retrieval-augmented generation. It analyzed LLM architectures and integration patterns and produced explanatory material showing how tokenization, transformer attention, and human-reviewed training shaped model behavior. The work documented how these components interact within
I explained LLM internals, showing how tokenization, transformer attention, and parameter updates from human-reviewed training affect predictions and model responses when applied to pharmaceutical documents.
I explained how an LLM tokenized prompts, used transformer architecture and attention to prioritize context, and described training from human-reviewed data that adjusted model parameters to improve predictions.