Quant & Market Technology Stack

• Python for systematic trading

  • FastAPI + Uvicorn to run a low‑latency tick and candle engine.
  • Fyers WebSocket API (fyers_apiv3.FyersDataSocket) for live NSE NIFTY futures market data.
  • Custom tick handler to convert exchange cumulative volume into true per‑tick volume.
  • 1‑minute OHLCV candle aggregation from raw ticks using exchange timestamps.
  • Real‑time order‑flow classification (high‑notional BUY / SELL / NEUTRAL) using next‑tick price reaction.
  • Business‑level volume threshold based on futures price × per‑tick volume (notional traded per impulse).

• Market microstructure & order‑flow modelling

  • Focus on aggressive flow (trades at/near bid‑ask) rather than just end‑of‑bar data.
  • Design of a live volume footprint / institutional prints feed via markers at exact trade prices.
  • Use of configurable notional thresholds to approximate institutional vs retail participation.

• Data engineering for trading

  • PostgreSQL (via psycopg2) for tick‑level storage of NIFTY futures (timestamps, LTP, per‑tick volume, bid/ask, totals).
  • Batched inserts for write‑optimized tick ingestion (MAX_BATCH_SIZE).
  • Schema suitable for later research, backtesting, and risk monitoring on top of the same live data.

• Real‑time infrastructure

  • WebSocket broadcast layer (FastAPI WebSocket + custom connection manager) for pushing:
    • 1‑minute OHLCV candles.
    • High‑notional BUY/SELL markers {time, price, volume, result}.
  • Health and monitoring endpoints (/, /health) exposing live server, database, and data‑feed status.

• 

  • 🎥 Live Backend Demo