Loading
Ace Intelligence Systems
Preparing a calmer, clearer view of your automation workspace.
Aligning workflows
Loading
Preparing a calmer, clearer view of your automation workspace.
Ace Intelligence
Production-ready automated trading platform combining ensemble strategies, LSTM prediction, and reinforcement learning — validated on 10,000 candles with $77K+ portfolio profit.
TradeLikeWhale is an enterprise-grade trading system that orchestrates multiple signal sources — technical indicators, neural networks, and a reinforcement learning agent — into a unified trading engine. The platform features a React dashboard for real-time monitoring, a FastAPI backend for strategy management, PostgreSQL for trade persistence, and direct MetaTrader 5 integration for live execution.
Clone the repo, install dependencies, configure your MT5 credentials, and launch the full stack with Docker Compose.
git clone https://github.com/OMCHOKSI108/TradeLikeWhale
cd TradeLikeWhale
pip install -r requirements.txt
cp .env.example .env
# Configure MT5_LOGIN, MT5_PASSWORD, MT5_SERVER in .env
python run.pyThe system is built as a four-layer pipeline: Data Ingestion (MT5 live feed, historical CSV, database cache) → Signal Generation (technical indicators, LSTM prediction, RL agent optimization, ensemble voting) → Risk Management (ATR sizing, drawdown limits, position validation) → Execution & Monitoring (MT5 order execution, trade logging, React dashboard, alert system). Each layer is decoupled and independently testable.
DATA LAYER
MT5 Live Feed ─┐
Historical CSV ─┼──> Data Processor
Database Cache ─┘ │
v
SIGNAL LAYER
Technical Indicators (ADX, RSI, MACD, BB) ─┐
LSTM Neural Network (price prediction) ───┼──> Ensemble Voting → Signal
RL Agent (adaptive optimization) ───┘
│
v
RISK LAYER
ATR Position Sizing ─┐
Daily Loss Limits ──┼──> Risk Check → Approved/Rejected
Drawdown Monitor ──┘
│
v
EXECUTION LAYER
MT5 Order Execution → Trade Logger → PostgreSQL
│
v
React Dashboard ← FastAPI ← Trade History
Alert System ← WebSocket ← Real-time P&LEnsemble voting combines four technical indicators (ADX for trend strength, Stochastic for momentum, RSI for overbought/oversold, MACD for trend direction, Bollinger Bands for volatility) with ML predictions. Each signal is weighted and voted into a final trade decision. The RL agent continuously adapts weights based on market regime changes.
LSTM neural networks are trained per symbol and timeframe to predict short-term price movements. A reinforcement learning agent (PPO-based) optimizes strategy parameters in real time, adjusting position sizing and signal thresholds as market conditions evolve. Models are retrained periodically using the latest market data to prevent drift.
ATR-based dynamic stop-loss and take-profit levels adjust automatically to market volatility. Daily loss limits halt trading after a configurable threshold. Drawdown protection monitors equity curves in real time and triggers an emergency stop if drawdown exceeds the configured limit. Position sizing supports fixed %, volatility-adjusted, and fixed USD methods.
Risk Parameters:
Daily loss limit: $500
Max drawdown: 10%
Position sizing: volatility_adjusted
Risk/reward ratio: 2.0
Max open positions: 5Walk-forward validation using 70% train / 30% test split on 10,000 historical candles per symbol. Best performers are USDJPY and BTCUSD on the 15-minute timeframe, showing consistent out-of-sample performance and positive Sharpe ratios across all tested instruments.
Symbol | Timeframe | Profit | Win Rate | Sharpe
USDJPY | 15min | $6,118 | 55.4% | 1.87
BTCUSD | 15min | $3,261 | 61.2% | 2.14
EURUSD | 4hr | $769 | 56.9% | 1.52
GBPUSD | 4hr | $231 | 53.8% | 1.28
─────────────────────────────────────────────
Total: $77,256.74 across all strategiesBuilt with React and Next.js, the dashboard provides real-time monitoring of balance, equity, P&L, and margin levels. Users can configure and activate strategies, view trade history with performance analytics, modify TP/SL on open positions, and receive alerts for trades, risks, and system events. Plotly charts display equity curves and symbol breakdowns.
FastAPI backend with JWT authentication exposes RESTful endpoints for strategy management, trade history, performance metrics, and system control. Full OpenAPI documentation is available at /docs. The API supports multi-user access with role-based permissions.
POST /token — Login
GET /api/status — Bot status & stats
POST /api/control — Start/stop bot
GET /api/positions — Active positions
GET /api/strategies — List strategies
POST /api/strategies — Create strategy
GET /api/history/trades — Trade historyPython powers the core trading engine with TensorFlow for LSTM models. FastAPI serves the REST API. React/Next.js renders the dashboard. MetaTrader 5 handles live execution. PostgreSQL persists all trade and strategy data. Docker Compose orchestrates the full stack deployment.
Python | FastAPI | React | Next.js | MetaTrader 5 | TensorFlow | PostgreSQL | Docker | Redis | StreamlitThe full source code is available on GitHub. Note: The repository is currently private to protect proprietary trading strategies and client data. Contact us for access or a demo.
https://github.com/OMCHOKSI108/TradeLikeWhale — Private (contact for access)Built by the Ace Intelligence founding team.
Om Choksi (CTO) — https://github.com/OMCHOKSI108
Yash Khare (Founder) — https://github.com/firefistisdead
Ansh Gajera (CEO) — https://github.com/anshgajera