Hybrid-RAG is a hybrid Retrieval-Augmented Generation (RAG) model that leverages BERT for retrieving relevant documents and GPT-2 for generating more accurate responses.

AI-powered Hybrid RAG system that combines dense semantic search and BM25 keyword search with Reciprocal Rank Fusion (RRF) to retrieve accurate answers from documents. Built with ChromaDB, BGE embeddings, and LLaMA 3.3 via Groq for fast, reliable, source-grounded responses.

The Open-Source Enterprise Meeting Intelligence Platform. A self-hosted, privacy-first alternative to Fireflies.ai & Otter.ai. Features autonomous recording (Zoom/Teams/Meet), Global RAG for cross-meeting chat, speaker diarization, and seamless 2-way sync with Jira & Slack. Built with Next.js 15, AWS Lambda, and OpenAI.

Search and Answering system over BITS Pilani research regulations and policies using Hybrid RAG

A hybrid RAG prototype combining retrieval and language model generation to answer queries over documents or knowledge sources.

This platform functions as a "bridge" between an organization’s internal operations and the strict requirements of international security and privacy frameworks

This project is a agentic AI framework with hybrid RAG system.

Role-aware medical RAG chatbot for MediAssist Hospital — hybrid vector search (dense + BM25), SQL RAG for structured claims data, and collection-level access control enforced at the Qdrant retrieval layer

A MSHA regulation AI agent. The system uses a novel RAG system - Advanced Parallel HybridRAG - Intelligent Fusion

Selbst-lernendes, selbst-tunendes KI-Gedächtnis auf Markdown + git — HybridRAG, echte Lern-Schleife, responsives Terminal-Dashboard (memDash). Fail-open, stdlib-Kern, MIT.