We build trading bots for IBKR, MT5 and TradingView.

Needed a market-neutral trading system — one exposed to the relationship between two stocks, not the direction of the market.

A 3-phase system: offline GICS-industry pair screening with a statistical test battery, a live execution engine with Kalman-filtered hedge ratios and automated news-risk blacklisting, and a monitoring dashboard.

Python 3.12 · Interactive Brokers · Databento market data · Kalman filtering · Google Gemini

Production system with guaranteed combo execution, no overnight risk, no martingale — backed by a 550-test automated suite before going live.

Needed fully automated SPY intraday execution with layered risk controls, not a signal bot that still requires a human to click buy.

An asyncio producer/consumer engine on the raw IBKR TWS API — decoupled tick ingestion and order execution, OCO entries, staged stop-loss/take-profit logic, automatic reversal on stop-out, and hard daily-loss and time-based flatten rules.

Python · asyncio · Interactive Brokers TWS API (EWrapper/EClient)

Runs unattended with hard PnL stops, trailing take-profit tiers, and a forced end-of-day flatten — no manual intervention required.

A trade-alert service publishes signals by email — needed those turned into verified, routed, risk-checked orders across multiple IBKR accounts, with a human approval step before anything trades.

An IMAP-polling signal parser handling two independent portfolio types, a Telegram approve/reject workflow, per-account capital allocation and position sizing, and dynamic multi-gateway routing — new IBKR accounts added via config, no code changes.

Python · ib_insync · Telegram Bot API · Playwright · SQLite

One signal feed reliably fanned out to multiple IBKR accounts with independent sizing and a human-in-the-loop approval gate, plus real fill confirmation instead of blind execution.