Multi-Timeframe Trend Recognition System

Multi-Timeframe Trend Recognition System Project Title: Development of a Multi-Timeframe Trend-Following Engine with Telegram Integration Background & Objective: We are transitioning from a short-term 15-minute trading signal model to a more advanced system that can recognize, track, and communicate multi-hour trend movements. The goal is to build a trend-following engine that understands market structure (A → B → C), maintains memory of past movements, and delivers more meaningful, context-aware trading insights through Telegram. Scope of Work: This is a new development phase, not a modification of the existing system. The proposed work includes: 1. Trend Classification Build machine learning classifiers for: 1-Hour 4-Hour Daily timeframes 2. Trend Memory Module Implement logic to track ongoing trends and remember direction/state across timeframes 3. Signal Fusion Engine Combine long-term trend direction with short-term movements to identify: Pullbacks vs. Reversals Trend Continuation opportunities 4. Telegram Message Layering Redesign output messages to reflect: Multi-timeframe trend bias (e.g., “1H Long Bias | 5m Pullback”) Momentum and structural changes in real time 5. Evaluation & Backtesting Develop robust backtesting logic to validate: Trend accuracy Signal stability Trade-worthiness 6. Deployment Integrate with the current infrastructure Live testing and performance monitoring Deliverables: Working trend classifiers (1H, 4H, Daily) Trend memory logic Signal fusion system Updated Telegram output system Backtesting reports Deployment-ready codebase Candidate Requirements: We’re looking to hire someone with expertise in: Machine Learning for time-series/trading data Python (NumPy, pandas, scikit-learn, PyTorch or TensorFlow) Signal processing and pattern detection Telegram bot integration Backtesting and model validation Experience in multi-timeframe trading logic is a strong plus