Currently an Extern @Pfizer·🟢 Open to work
I explore AI-driven document insights and data extraction, building practical skills through a Pfizer externship.
Portfolio items highlight work from the Pfizer externship: document analysis, data extraction pipelines, and AI-driven insight generation.
Designed and developed an AI powered automation platform that streamlined developer workflows across Windows, macOS, and Linux. Reduced environment setup time from 40 minutes to 2 minutes through workflow automation. Integrated OpenAI APIs to automate repetitive processes through natural language
Led the development of an AI powered career guidance platform serving 1,000+ students. Built backend services using Django REST Framework, Python, SQL, OpenAI GPT 4, and Google APIs. Integrated OpenAI GPT 4 and Google APIs to generate personalised recommendations for users. Worked closely with users
I explore AI-driven document insights and data extraction, building practical skills through a Pfizer externship.
I am Martins Ejike. I have little work experience and am at the start of my career. I completed the Pfizer Advanced: AI-Powered Document Insights & Data Extraction externship, where I worked on using AI for document analysis and data extraction.
Externships
Pfizer Advanced: AI-Powered Document Insights & Data Extraction Externship
Pfizer
Experience
Software Engineer Intern
Irish Life, Dublin · Jan 2023 – Sep 2023
Student Activator (Operations & Digital Transformation)
Dublin City University, Dublin · Sep 2022 – Jan 2023
Education
University College Dublin
MSc Computer Science (Artificial Intelligence) · Class of 2025
Dublin City University
BSc Enterprise Computing · Class of 2024
Skills
Designed and developed an AI powered automation platform that streamlined developer workflows across Windows, macOS, and Linux. Reduced environment setup time from 40 minutes to 2 minutes through workflow automation. Integrated OpenAI APIs to automate repetitive processes through natural language
Designed and developed an AI powered automation platform that streamlined developer workflows across Windows, macOS, and Linux. Reduced environment setup time from 40 minutes to 2 minutes through workflow automation. Integrated OpenAI APIs to automate repetitive processes through natural language interactions. Built Python automation services using subprocess and operating system interactions.
Led the development of an AI powered career guidance platform serving 1,000+ students. Built backend services using Django REST Framework, Python, SQL, OpenAI GPT 4, and Google APIs. Integrated OpenAI GPT 4 and Google APIs to generate personalised recommendations for users. Worked closely with users
Led a team of five to win the Engineer's Choice Award by building an AI powered language learning application in 48 hours. Developed AI powered text to speech functionality for low resource African languages using OpenAI technologies.
✅ 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-powered document intelligence for enterprise PDFs, applying OCR, retrieval-augmented generation, and extraction techniques. The work analyzed LLM functionality and documented how model components and retrieval layers interact. Deliverables included technical explanations and design notes on model behavior and integration.
.jpg)
I produced a technical explanation showing how LLMs tokenize text, use Transformer attention to model word relationships, and how pretraining followed by curated fine-tuning and human feedback improves model accuracy and safety.
I explained how LLMs tokenized text and used Transformer architectures with attention to model word relationships. I described pretraining on large corpora and subsequent fine-tuning with curated data and human feedback to improve accuracy and safety.
During the externship I cleaned datasets with Pandas, parsed and flattened nested JSON into a DataFrame, performed text standardisation (lowercasing, punctuation removal, encoding fixes), and applied denoising, CLAHE, and adaptive thresholding to enhance scanned documents for OCR.
I processed an SDF with PyMuPDF to extract words and bounding boxes, used regex and anchor phrases to link labels to dates, and reconstructed table content. The deliverable documented extracted fields, layout challenges, and recommended next steps.