In Progress
RAG Playground
An interactive sandbox for testing retrieval-augmented generation strategies — chunking sizes, retrievers, and re-ranking approaches — against a shared document set.
RAGVector SearchLLMs
Experiments, prototypes, and in-progress AI tooling — the workbench behind the finished projects.
An interactive sandbox for testing retrieval-augmented generation strategies — chunking sizes, retrievers, and re-ranking approaches — against a shared document set.
An autonomous agent that crawls a site, evaluates technical SEO signals, and produces a prioritized fix list — the same engine that powers this portfolio's SEO discipline.
A catalog of reusable prompt patterns for common LLM tasks (extraction, classification, tool use) with notes on when each pattern breaks down.
A small tool that visualizes how a piece of text is tokenized across different model families, useful for building intuition about cost and context limits.