01 Problem
Indian retail investors had no screener combining fundamental, technical, and risk analysis with explainable ML. Existing tools were basic filters or black-box tips.
02 Approach
Three-service architecture: (1) FastAPI quant service — 5 screening strategies, multi-agent analysis with scikit-learn + networkx, yfinance data; (2) NestJS/MongoDB backend — auth, paper trading, portfolio, intelligent caching; (3) Next.js 16 frontend — Recharts, TanStack Query, Apple-inspired design, PWA. Cloudflare tunnel for remote access. Paper trading falls back to localStorage.
03 Outcome
Screener ranking ~200 NSE stocks with combined/fundamental/technical/risk scores (0–10). Multi-agent analysis explains scoring. Paper trading + portfolio. PWA installable on iOS. Quant service caches intelligently for yfinance rate limits.
04 Why it matters
Because an everyday investor's money deserves the same seriousness as an institution's.