An AI assisted EA forex watch-list scanner (not an auto-trader) of PDL/PDH bounces on M15 is needed.

A developer will create an AI assisted EA forex watch-list scanner (not an auto-trader) designed to surface PDL/PDH bounces on M15, with coding in Python/MQL5. This will strictly be a market intelligence and scanning tool (no automatic order placement) and EA primary function will be a delivery of visual dashboard UI, 2-phased alerts (warning vs. execution), and multi-channel notification.

1. Project Overview & Objective

I am seeking an experienced MQL5 and Python developer to build a professional, non-auto-trading Multi-Symbol, Multi-Timeframe (MTF) Watch-list Scanner Expert Advisor (EA) for MetaTrader 5.

Unlike standard indicators, this system utilizes a hybrid architecture, leveraging MQL5 for native MT5 data extraction and macro-timeframe bias filters, while utilizing a Python-generated .onnx machine learning model to handle complex micro-timeframe price action recognition and EMA ribbon dynamics.

• Primary Function: Visual dashboard UI, phased alerts (warning vs. execution), and multi-channel notifications.
• Non-Goal: Zero automated order placement. This is strictly a market intelligence and scanning tool.

2. Core Technical Architecture & Tech Stack

• Platform: MetaTrader 5 (MQL5)
• AI Integration: Native MQL5 ONNX runtime execution utilizing a custom pre-trained model (.onnx file format) developed in Python.
• Data Flow: MQL5 extracts multi-timeframe price and indicator data, formats arrays, passes them to the .onnx inference engine, and reads back structured pattern/phase signals.

3. Scope of Work & Key Modules Required

Module A: Multi-Symbol & Multi-Timeframe (MTF) Engine (MQL5)

• Master watch- input list of 29 pairs  (e.g., EURUSD, GBPUSD, AUDUSD) with auto-sync to MT5 Market Watch.
• Macro Bias (  / ) & Micro Bias ( ): Calculates structural filters and trend states directly in MQL5 to govern overarching trade direction.

Module B: AI-Driven Micro Engine & Pattern Recognition (Python / ONNX)

• Key Level Tracking: Computes Previous Day Low (PDL for Buy setups) and Previous Day High (PDH for Sell setups).
• Pattern Library: Integration of an ONNX model capable of identifying extended reversal formations (hammers, pin bars, engulfing, spinning tops) under various touch/pierce variations.
• EMA Ribbon Dynamics: Monitors M15 EMA ribbon compression phases and detects trend inflexion points.

Module C: Phased Alert & Warning System

• Phase 1 (The Warning): Fires a low-priority alert when price enters a customizable proximity zone (e.g., within 10 pips of PDL/PDH) while the M15 EMA ribbon is actively compressing.
• Phase 2 (The Signal): Fires a high-priority trade entry signal exclusively when a confirmation candle closes in alignment with the Phase 5 ribbon inflexion and MQL5 macro/micro bias.
• Adjustable Delay: Configurable confirmation candle delay parameter (  to candles).
• Notification Channels: Native MT5 Popup, Push Notifications (Mobile), and Email alerts with built-in anti-spam duplicate suppression.

Module D: Interactive Dashboard UI (MQL5)

• A clean, non-intrusive on-screen panel grid displaying monitored symbols.
• Columns: Symbol, Macro/Micro Bias, Ribbon State, Distance to Key Level, and Actionable Status (Approaching / Compressing vs. Signal Confirmed).

4. Developer Deliverables

1. Source Code: Well-documented .mq5 and .mqh files, plus the accompanying Python training/export script used to generate the .onnx model.
2. Compiled Binaries: Ready-to-use .ex5 file and the corresponding .onnx file placed in the correct MT5 directories (MQL5\Files).
3. Setup Documentation: Brief instructions on how to load the .onnx file, compile the EA, and configure inputs.

5. Ideal Candidate Requirements

• Proven track record of developing advanced MQL5 Expert Advisors and Custom Indicators.
• Explicit, demonstrable experience integrating Python machine learning models (.onnx) into MT5.
• Strong background in building complex, non-blocking visual UI dashboards in MQL5.
• Ability to sign a mutual Non-Disclosure Agreement (NDA) before viewing proprietary project details.

If you are qualified for this hybrid MQL5/Python build, please reply with examples of past MT5 projects involving custom indicators, ONNX models, or complex dashboard interfaces.