I already have several Python trading strategies coded and tested in isolation. What I need now is a clean, production-ready pipeline that does three things seamlessly: lets me back-test those strategies on historical data, pushes selected ones live through Zerodha Kite, and keeps an eye on real-time performance. The stack is non-negotiable: Pandas and NumPy for data wrangling, Backtrader as the core engine, and Zerodha’s Kite Connect/WebSocket API for brokerage connectivity. Wherever latency can be shaved, I’d like to see Numba or Cython used for just-in-time speed-ups, especially inside tight loops. The live leg must subscribe to tick data over WebSockets, place/modify orders, and stream portfolio updates without choking. Deliverables • Unified codebase that plugs my existing strategy classes into Backtrader • Back-testing module (CLI or notebook) that outputs detailed metrics and equity curves • Live-trading runner wired to Kite, with basic risk controls and logging • Lightweight performance monitor (console or simple dashboard) showing PnL, drawdown, and order status in real time • Setup/usage guide so I can spin everything up on another machine in minutes Acceptance will be based on: 1. Accurate replication of my strategy results in the new back-test module. 2. Successful paper-trade session on Kite with real-time metrics visible. 3. Code quality, clear structure, and concise documentation. If you’re comfortable juggling Backtrader, Kite WebSockets, and a bit of low-level optimization, let’s talk.