Algo developer — I build the systems that trading runs on.
Live market data for Indian equities, F&O and options: broker WebSocket feeds, tick-to-bar engines, option-chain resolution, and the dashboards that make all of it legible in real time. Most of what I write connects at 09:15 and has to still be right at 15:30.
📍 Bhavnagar, India · 💼 Open to algo developer / trading systems roles · 🌐 tejgohel.github.io · ✉️ tejgohel.tg18@gmail.com · 🔗 LinkedIn
Live data infrastructure. Broker market feeds over WebSocket, binary packet parsing, tick-to-candle construction, session-anchored VWAP, incremental indicator state that updates in microseconds instead of recomputing history on every print. Roughly 200–500 instruments streaming at once, with the socket thread doing nothing but writing memory.
Instrument and universe resolution. F&O universes rebuilt daily from the exchange instrument master; option strike grids detected per stock from real chain gaps rather than assumed; expiry rolls; liquidity floors. This is the unglamorous half of options tooling and it is where most implementations quietly break.
Correctness under market conditions. A surprising amount of this work is
knowing which broker field lies and when. A net_change that is always zero.
An ohlc.close that silently flips to today's close after the session ends.
A Saturday run where "today" means Friday, producing a perfectly plausible and
completely flat heatmap. Each of those is a real bug I shipped, found, and now
guard against explicitly.
Measuring before shipping. Multi-year historical studies (80k–110k stock-days) behind anything that claims an edge, because a rule that has not been measured over a meaningful sample is a guess with extra steps. Currently going deeper on the quantitative side of this.
| Project | What it is |
|---|---|
| nse-algo-bot | Intraday trading bot for NSE equities — market feed, per-tick evaluation, leverage-aware sizing, order placement, and stop / trail / profit-lock / target management through to a forced square-off. Copy trading built in |
| OPTION-DASHBOARD | Options intelligence platform — live option chain, futures and equity analytics, 200+ derived metrics per strike, regime detection, self-scoring signal engine. FastAPI · React · TimescaleDB · Redis · TypeScript |
| nse-live-heatmap | Live treemap of the entire NSE F&O universe, repainting from the market feed. Universe resolved from the exchange master, previous close verified before a single tile is painted |
| option-strike-scanner | Resolves a tradable strike per F&O stock before the open, then scans the option's own 5-min price live. Strike-grid auto-detection, tick-to-bar building, session VWAP |
| smc-ob-scanner | Multi-timeframe order-block scanner for NSE equities and Forex — a TradingView Pine indicator ported to Python, plus a Flask dashboard, CLI and MQL5 expert advisor. Ships with synthetic data so it runs with no broker account |
| crypto-signal-scanner | Binance spot scanner — WebSocket engine, SQLite candle store, SSE dashboard, Telegram alerts |
Several of these ship the full engine with the signal rule left as a documented plug-in point. That is deliberate: the infrastructure is the reusable part and the part worth reviewing, and live trading rules are not something anyone publishes. Every repository runs end to end once a rule is supplied — or, in
smc-ob-scanner, straight out of the box on bundled data.nse-algo-botis the one that can place real orders; it ships with paper trading on by default.
Core
Services & storage
Frontend
Markets
Only nse-algo-bot can place orders, and only once you write a rule for it
and turn paper trading off. None of this is investment advice.