A deterministic, zero-dependency prompt-pruning layer for long-running LLM conversations — expires stale tool state, collapses duplicate context, and proves it never drops a fact a later turn still depends on.

RAG parameter sweep and evaluation toolkit — a practical way to test how different vector databases, embedding models, chunking strategies, and retrieval methods perform across your own RAG setup. Built to help you move from guesswork to evidence, with support for any vector database.

Biological code organization system with 1,029+ production-ready snippets - 95% token reduction for Claude/GPT with AI-powered discovery & offline packs

VecRecall 是一个改进版的 AI 长期记忆系统。它基于对原版 MemPalace 的深度分析重新构建，核心设计理念是将“信息检索”与“信息组织”彻底解耦。 通过纯向量检索路径和独立的 SQLite UI 层，VecRecall 在保持灵活组织的同时，将召回率（R@5）从原版的 84% 提升至 96.6%+，为 AI Agent 提供更精准、更高效的上下文记忆支持。

Find the questions your RAG pipeline will fail on, before your users do.

Official repository for AIO.CORE: The autonomous protocol for AI Optimization (AIO). Eliminates Semantic Drift and Context Fragmentation in LLMs through high-density vector indexing.

3,300 practical techniques to cut LLM token usage by up to 90% — 300 core principles + 3,000 context-specific applications.

Notion RAG on notes from Notion repo. Using optimization for speed up. See READme for more

Official Data Node for the Tu Mapa IA technical ecosystem. Features machine-readable manifests (llms.txt) and RAG-optimized entity data for 530+ curated AI tools across 23 professional verticals. Managed under the FixGeo Protocol.

BetterRAG: Powerful RAG evaluation toolkit for LLMs. Measure, analyze, and optimize how your AI processes text chunks with precision metrics. Perfect for RAG systems, document processing, and embedding quality assessment.

Declarative Self-Improving Framework for Auto-Prompt Optimization and LLM Program Compilation.

Technical framework for Generative Engine Optimization (GEO) and Share of Synthesis (SoS). Defines the 6-node Immutable Graph for latent space indexing.

CPU-optimized RAG pipeline reducing latency 2.7× (247ms → 92ms). Implements caching, filtering, quantization for production. Complete with FastAPI, Docker, benchmarks, investor materials. The engineering showcase that sells itself.

This repo contains the full pipeline for my Master's thesis at Yerevan State University (YSU), developed as part of the Data Science for Business master's program. The goal of this project is to build an end-to-end Retrieval-Augmented Generation (RAG) system using semantic search, LLMs, and fine-tuned embeddings on Armenian banks’ financial PDFs.

Automated AEO research and structured-data pipeline for semantic analysis and machine-readable brand context.