BigviewBot Trading Engine

An automated, full-stack engine algorithmic trading engine hosted on AWS EC2, integrated with an Alexa+ skill interface and a real-time React monitoring terminal.

Comment

📌 Inspiration

Algorithmic trading moves fast, but managing complex, automated trading bots traditionally requires staring at terminal logs, monitoring complex web dashboards, or scrambling for a laptop when market conditions rapidly pivot. We asked ourselves: What if a quantitative trader could monitor, control, and execute risk management across an entire trading ecosystem using natural voice commands?

We set out to build a seamless bridge between modern AI voice interfaces (Amazon Alexa) and real-time futures execution on WEEX via CCXT. Whether checking portfolio metrics while away from the desk or halting trading instantly in volatile markets, our solution makes multi-engine quant management instantaneous and completely hands-free.

⚙️ What It Does

Our system is a full-stack multi-engine quant control center featuring direct voice integration through Amazon Alexa:

  • Voice-Activated Portfolio Analytics: Real-time query of total portfolio valuation, available USDT margins, active position distributions, and win rates via GetPortfolioBalanceIntent.
  • Live CCXT Asset & Market Tracking: Fetch real-time ticker prices and 24-hour percentage performance metrics directly from WEEX swap markets (GetAssetPriceIntent & GetAssetPerformanceIntent).
  • Multi-Turn Emergency Control Center: Say "Alexa, pause trading" to trigger a interactive voice confirmation loop (ToggleTradingEngineIntent + ConfirmActionIntentHandler). Confirming with "Yes" triggers an instant global Emergency Kill Switch across all 5 automated trading loops.
  • Rich Visual Feedback (APL): Screen-enabled Alexa devices (Echo Show & Developer Simulators) dynamically render live dark-mode dashboard cards showing primary financial metrics, active engine counts, and system status indicator colors (#00FF66 active / #FF3333 paused).
  • Real-time Engine Activity & Audit Logs: Instantly inspect active trades, recent execution history, and raw system logs straight from memory (GetEngineLogsIntent).

🛠️ How We Built It

Technical Architecture ┌─────────────────────────┐ ┌──────────────────────────────┐ │ Amazon Alexa Voice │ ───► │ Express.js Server │ │ (ASK SDK + Custom APL) │ │ (ask-sdk-express-adapter) │ └─────────────────────────┘ └──────────────┬───────────────┘ │ ▼ ┌─────────────────────────┐ ┌──────────────────────────────┐ │ WEEX Exchange │ ◄───► │ CCXT Library Integration │ │ (Live Futures / Swaps) │ │ (Real-time Market Tickers) │ └─────────────────────────┘ └──────────────────────────────┘

  • Backend Stack: Node.js, TypeScript, Express.js.
  • Voice Pipeline: Alexa Skills Kit (ask-sdk-core, ask-sdk-express-adapter) using custom Intent Handlers for state management and intent routing.
  • Confirmation Flow: Utilized Alexa Session Attributes to handle context-aware multi-turn dialogues (AMAZON.YesIntent / AMAZON.NoIntent) before executing system-wide mutations.
  • Exchange Integration: Integrated ccxt.weex with strict symbol normalization to handle voice slot variations (e.g., mapping "Bitcoin" \rightarrow BTC/USDT).
  • Visual Interface: Alexa Presentation Language (APL 2023.3) for responsive data visual card rendering on smart displays.

🔬 Challenges We Overcame

  • Multi-Turn State Persistence in Voice: Implementing a fail-safe confirmation flow required strict session attribute mapping (pendingAction = 'TOGGLE_KILL_SWITCH') to ensure accidental voice triggers wouldn't inadvertently freeze active market positions.
  • Slot Normalization for Voice Commands: Spoken crypto names vary dramatically. We built a robust symbol normalization helper mapping phonetic inputs like "Solana" or "Ethereum" directly into valid WEEX USDT contract pairs (SOL/USDT, ETH/USDT).
  • Graceful Screen Compatibility: Ensuring the Alexa handler conditionally injects APL visual directives only when the target hardware supports screen interfaces (supportsAPL check), preventing execution crashes on audio-only Echo hardware.

🏆 Accomplishments That We're Proud Of

  • Sub-Second Execution: Achieved near-instant voice recognition to market action pipeline latency for emergency risk control.
  • Flawless Multi-Turn Dialogue: Implemented a clean confirmation workflow that safely intercepts destructive voice commands before mutating live global server states.
  • Rich APL Experience: Built dark-mode Echo Show visual cards that complement voice output with real-time color-coded metrics.

📖 What We Learned

  • How to structure complex ASK SDK Request Handlers in TypeScript using sequential evaluation ordering.
  • Best practices for combining real-time REST API exchange connections with asynchronous voice request pipelines.
  • Designing multi-modal UI/UX where audio speech and visual screen layouts reinforce each other without overwhelming the user.

🔮 What's Next for BigviewBot trading engine

  • Voice-Gated Order Execution: Introduce customizable limit/market order placement with voice PIN confirmation.
  • Push Notification Alerts: Leverage Alexa Event Gateway to proactively notify users when stop-loss or take-profit targets are triggered on WEEX.
  • Custom Quantitative Indicator Queries: Allow traders to ask for real-time technical indicator state (e.g., "Alexa, what is the 14-period RSI for BTC on WEEX?").
Share this project:

Updates

Submission history