Back to Projects
In Progressai-ml

Trading RAG Pipeline

Local RAG system for finance and trading knowledge management with semantic search and optional LLM-powered synthesis.

1 min read
Solo project
PythonFastAPIQdrantSupabasen8nOllamaDocker
View Code

The Problem

Trading knowledge is scattered across YouTube videos, articles, and notes. Finding relevant insights requires manually searching through hours of content.

The Solution

Built a local RAG pipeline that ingests content via n8n workflows, chunks documents with token-aware processing, and enables semantic search with optional LLM synthesis.

Key Features

  • Document ingestion from multiple sources
  • Token-aware chunking with timestamp preservation
  • Vector-based semantic search with metadata filtering
  • Backtest parameter tuning with grid/random search
  • Admin dashboard with leaderboards