Portfolio Site
Building AI systems that scale.
The problem is no longer adding features, but controlling the complexity. That is the work I like: understanding the model, building the system around it, and getting the result to survive contact with real users. These projects cover that work, from a transformer implemented from scratch to a resume generator with real backend, auth, and PDF plumbing to a RAG chatbot grounded on its own corpus. My background in analysis and enterprise integration turns out to be unusually useful here: ambiguous requirements, messy systems, and the human problem of figuring out what actually needs to be built.
Fundamentals of large language models and deep learning, including transformer implementation in Pytorch, training and inference, performance analysis, hyperparameter tuning, and architectural variations (ablations).
A serverless resume generator utilizing the Anthropic messages API to create a structured markdown/html output that is rendered into a polished PDF resume, tailored to the input job description.
A review of the site creation process and the design and integration of a RAG-powered Generative AI Chatbot