Pfizer Externship via Extern
I completed the Pfizer Advanced externship on AI-powered document insights and data extraction and I am growing practical experience in applied AI projects.
Portfolio of applied AI and data extraction work from my Pfizer Advanced externship and other hands-on projects that show practical skills and learning.
.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.
I completed the Pfizer Advanced externship on AI-powered document insights and data extraction and I am growing practical experience in applied AI projects.
I am Sultan Fohpa. I have some work experience and I completed the Pfizer Advanced: AI-Powered Document Insights & Data Extraction externship. I am early in my career and building practical skills through project-based work.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
In progress
Experience
ABM Developer
Dissertation Project · October 2025 – March 2026
Club Secretary
KCL Regents (American Football) · September 2025 – July 2026
AI Research Assistant
University Research · June 2024 – August 2024
App Developer
University Project · January 2024 – March 2024
Education
King’s College London
Bachelor of Computer Science with Honours · Class of 2026
Handsworth Grammar School for Boys
A-levels · Class of 2022
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 prototyped an AI-powered document-insights pipeline that combined OCR, LLM processing, and retrieval-augmented generation to extract structured data from PDFs. The project documented model choices and integration steps, and produced explainable notes on LLM behavior and tradeoffs observed during experiments.
I produced a written summary that traced an LLM pipeline end to end, explaining training on large datasets, prompt-to-embedding conversion, transformer context capture, and token-sequence generation.
I summarized how LLMs were trained on large datasets, converted prompts into vector embeddings, used transformer layers to capture context, and generated responses by selecting statistically likely token sequences, showing the model pipeline end to end.
