Notes

Working notes

Machine learning experiments, build write-ups, and opinions about where AI actually helps a business. Some of it is for engineers, some for owners.

In Generative AI by Robert Jean Pierre · Sep 3

GraphRAG vs RAG: When Vector Search Isn't Enough

GraphRAG promises to fix RAG's biggest weakness, but the 2025 benchmarks tell a sharper story. When a knowledge graph beats plain vector search, and when it doesn't.

15 min read

In Data Science by Robert Jean Pierre · Sep 2

When 95% Accuracy Isn't Enough: Detecting Violent Crime in Washington, D.C.

An end-to-end machine learning project using Washington, D.C. crime data to explore patterns, compare classifiers, and improve violent-crime detection.

10 min read

Get the notes by email

New posts when they go up. You confirm first, you can leave in one click, and the list is never sold, shared, or used for anything else.