Writing / topic

#LLM

  1. Why RAG Shouldn't Retrieve Everything

    In BARA, retrieval is three separate decisions — whether, from where and how much, then which chunks — made by deterministic rules before a vector is ever compared. A research note with the code.

    • AI
    • RAG
    • LLM
Personal / Open Source

BARA — Layered Memory Architecture

Behavior-adaptive retrieval for stateful conversational AI — four memory tiers behind 41 deterministic decision gates instead of always-on RAG.

  • Python
  • FastAPI
  • PostgreSQL
  • pgvector
  • NumPy
  • +6
◧ motion diagram
Personal / Open Source

SonicMirror

Spotify listening data turned into an acoustic fingerprint and run through a multi-stage Gemini prompt chain that produces a specific, falsifiable personality profile.

  • Next.js
  • React
  • TypeScript
  • Spotify Web API
  • Google Gemini
  • +2

Jump to