Currently an Extern @Pfizer
Sophomore at Saint Peter's University studying Data Science, gaining hands-on experience via a Pfizer externship in AI-powered document insights and data extraction.
Portfolio of coursework and work from a Pfizer externship, focused on AI methods for parsing documents and extracting data, plus class projects from my Data Science program.
.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.
Sophomore at Saint Peter's University studying Data Science, gaining hands-on experience via a Pfizer externship in AI-powered document insights and data extraction.
I am Labannya Debnath, a sophomore at Saint Peter's University pursuing a Bachelors in Data Science. I have limited formal work experience and I am building practical skills through coursework and a Pfizer externship focused on AI document insights and data extraction.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
Pfizer
Experience
Research Assistant
Saint Peter's Unversity · May 2025 - Aug 2025
Education
Saint Peter's University
BSc In Data Science · Class of 2028
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 externship explored AI-driven document intelligence for enterprise PDF workflows, combining optical character recognition, large language models, and retrieval-augmented generation. The work surveyed foundational LLM concepts and produced explanatory material that connected model mechanics to document-processing patterns.
I produced a focused summary of how large language models learn from corpora, use Transformer self-attention for context, generate text via next-token prediction, and are refined with human feedback for safer, more relevant outputs.
I summarized how large language models learned from large text corpora, used Transformer self-attention to model context, generated responses by next-token prediction, and were refined with human feedback to improve safety and relevance.