Currently an Extern @Pfizer
Masters student at USC in Applied Analytics, building practical technical skills with experience in teaching AI and data analytics as a Program Assistant at Stanford CARE.
Portfolio of coursework and externship deliverables related to operations strategy, people analytics, and data analysis tools used during my masters training.
.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.
Masters student at USC in Applied Analytics, building practical technical skills with experience in teaching AI and data analytics as a Program Assistant at Stanford CARE.
I am Brett Loy, a graduate student at the University of Southern California studying toward a master's degree. I am focused on data analytics and building technical and professional skills. I have completed an Amazon Operational Strategy & People Analytics externship as part of my training.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
Pfizer
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 project prototyped AI-powered document intelligence using OCR, LLMs, and retrieval-augmented generation to process enterprise PDFs. The work analyzed LLM mechanics, evaluated model behaviors, and documented tradeoffs relevant to building a pipeline for automated document extraction and question answering.
I explained that LLMs function as next-word predictors using transformer attention, and documented how human feedback and reinforcement learning refine their responses.
I explained that LLMs operate as next-word prediction models using transformer attention to weight relevant tokens, and that newer models used human feedback and reinforcement learning to refine response quality.
In the externship project I set up and ran Python notebooks on Colab, cleaned tabular data with Pandas, flattened nested JSON into a usable DataFrame, standardized text (lowercasing, punctuation removal, expansion), and applied denoising and contrast filters to improve OCR input.
I tested PyMuPDF on multi-page SDFs, extracted text including table content via bounding boxes, verified accurate text capture, documented layout challenges (none found), and recommended normalizing date formats for downstream field extraction.
