ML: Techniques Corpus

4 documents filed under ml. Every one of them is retrievable by agents through rag_context. All categories.

• From Naive RAG to Context Lakes: A Maturity Ladder for LLM Context (Hybrid Retrieval, GraphRAG, Code Graphs, Agent Memory)

  A rung-by-rung ladder for LLM context as of 2026-10: no retrieval, naive RAG, hybrid, query-side and index-side enrichment, agentic retrieval, GraphRAG, code graphs, agent memory and context lakes, assembly and evaluation, each with sourced gains, costs and a climb signal.

  Researched 2026-10-05 · 83 sources

• GhostNet: Efficient CNN Architecture Guide

  A comprehensive guide to efficient CNN architectures using ghost operations for mobile and edge deployment.

• RAG Quick Start with SQLite Vectors on Turso

  Add small, fast RAG to any codebase with zero extra services using libSQL/Turso vectors and an MCP tool.

• RAG Retrieval Quality: Hybrid Fusion and Chunking for Code Corpora

  Evidence-based guidance on hybrid retrieval fusion (RRF vs convex combination, with the counter-narrative), pg_trgm's structural failure as a lexical arm over long chunks (empirically verified), AST-aware code chunking, token-based sizing, contextual retrieval economics, and how to evaluate any of it.

  Researched 2026-07-26 · 16 sources