# Embedded Thought

> Systems programming, performance engineering, and the occasional deep dive.

Personal technical blog by Scott Everitt.

## Topics
- Vector databases and approximate nearest neighbor search
- Quantization algorithms (TurboQuant, Lloyd-Max, scalar quantization)
- Go systems programming and performance engineering
- Embedding models and semantic search
- File format design and memory-mapped I/O

## Posts

- [Building a Vector Database That Never Decompresses Your Vectors](https://scotteveritt.github.io/blog/quantization-native-vector-database/): How I built tqdb, a pure-Go embeddable vector database that stores vectors in 4-bit quantized form and searches them without decompressing. Honest benchmarks across four datasets, from 91.9% recall...


## Links
- Blog: https://scotteveritt.github.io
- RSS: https://scotteveritt.github.io/feed.xml
- Sitemap: https://scotteveritt.github.io/sitemap.xml