A multi-modal RAG pipeline with adaptive chunking, hybrid search, and LLM-driven re-ranking. Built to answer complex queries over large document corpora with high fidelity.
Projects
Some side quests for showcase.
PythonLangChainFAISSFastAPI
Jun 2025
Aug 2024
Lightweight chat interface over local LLMs. Supports streaming, system prompt injection, conversation branching, and model hot-swapping without page reload.
ReactFastAPIOllamaWebSockets
Nov 2024
Mar 2025
Synapse
Graph-based knowledge management system. Nodes represent concepts, edges encode relationships. Built for researchers who think in networks, not lists.
TypeScriptD3.jsNeo4jExpress
Apr 2024