All Documents
3,528 documents available
QWEN.md - MediaWiki Graph Analysis Platform
Documents architecture, setup, configuration, and conventions for a MediaWiki graph analysis toolkit using embeddings and clustering.
plan
Outlines a full Dockerized system that builds a knowledge graph from uploaded data, retrieves relational paths via PathRAG, and generates answers using an Ollama-hosted LLM.
GenAI Project Context
Defines a Go library that wraps Gemini, OpenAI, and Ollama under a single interface with built-in tools and retry logic.
Fields
Lists papers on counterfactual synthesis and bias elimination in machine learning, grouped by causal inference tools.
📝 Development Notes
Documents bugs, design decisions, and agent flow for a chatbot backend using Gemini, Pinecone, and a plugin system.
Multi-class: exactly one of the sentiment labels applies
Describes LabelFusion, a Python package that learns to combine a transformer classifier with one or more LLMs for multi-class and multi-label text classification.
EmergentDB Setup Guide
Walks through local setup of EmergentDB with document ingestion, Gemini OCR, and a Next.js frontend.
Qdrant & KiloCode Integration FAQ
Explains RAG, Qdrant collections, vectors, and the KiloCode workflow for local semantic code search.
HeadKey Memory System - Developer Cheat Sheet
Documents a memory API system for AI agents built on the CIBFE architecture, covering setup, configuration, and usage.
Kamibiki Implementation Plan
Plans a parallel implementation of a local git repository search engine using tree-sitter chunking and binary quantized embeddings.
Setup Guide
Walks through cloning, installing dependencies, setting up Ollama, and choosing between a FAISS or Cosmos DB RAG implementation.
🗺️ ModelAtlas Development Plan
Guides phased development of a multi-agent system for discovering, evaluating, and enriching LLM models compatible with Ollama.
Wiki Search - Technical Documentation
Documents a five-stage pipeline that loads Wikipedia articles, chunks them, generates 384-dim embeddings, builds a FAISS index, and runs interactive semantic search.
alligator.ai - Agentic Legal Research Platform
Defines product requirements for an AI-powered legal research platform that automates case analysis and strategy for boutique litigation firms.
draft
Lists and briefly describes transformer-based computer vision papers, from standard architectures to few-shot learning, with diagrams and code links.
index
Documents a project that scratch-trains tiny GPT-2 language models on synthetic news bulletins, achieving ~80% of Llama 8B quality with 300x fewer parameters.
Vector RAG POC - Frequently Asked Questions
Answers 25 common questions about setting up and using a Vector RAG proof of concept with Elasticsearch.
MLOps Guide: MLFlow, DVC & Evidently
Shows how to combine MLFlow, DVC, and Evidently for experiment tracking, data versioning, and model monitoring in a Python project.
Carapace Technical Design Specification (SPEC)
Defines a triage orchestration engine that detects similar pull requests and issues, selects canonical candidates, and routes noise away from maintainer views.
RFC-BLite: High-Performance Embedded Document Database for .NET
Specifies an embedded document database for.NET with zero-allocation I/O, C-BSON format, and ACID transactions.
PRD: GNO - Local Knowledge Index and Retrieval (CLI + MCP)
Defines a local-first knowledge indexing and retrieval system with CLI and MCP interfaces, using SQLite FTS5, vector search, and local LLM inference.
Run Phase - Visual Product Search
Describes how to implement visual product search using on-device CLIP embeddings and Couchbase Lite vector similarity queries.
Current Status: v0.4.0 In Development
Tracks completed and upcoming work for v0.4.0 of a memory graph library with HiRAG, sync, and re-ranking features.
Criterion MCP Server
Documents a hosted MCP server that exposes three semantic search tools for Quran verses and Hadith collections to AI assistants.