Development of a Scalable AI-Powered Predictive Maintenance System for Industrial Machinery using LSTM Networks

We are seeking an experienced AI engineer to develop a predictive maintenance system for our fleet of industrial machinery. The system should leverage historical sensor data (temperature, pressure, vibration) to predict equipment failures and schedule maintenance proactively. The core of the system will utilize Long Short-Term Memory (LSTM) networks implemented in Python with TensorFlow or PyTorch. Data preprocessing will be crucial, involving outlier detection, missing value imputation, and feature scaling to ensure model accuracy. The solution must be scalable to handle a growing number of machines and sensors, utilizing cloud-based infrastructure (AWS or Azure) for data storage and model deployment. We require a robust solution with clear documentation, automated training pipelines, real-time prediction capabilities, and an intuitive dashboard for visualizing maintenance schedules and equipment health. Experience with time series data analysis, anomaly detection techniques, and model evaluation metrics (e.g., precision, recall, F1-score) is essential. The system must integrate with our existing CMMS (Computerized Maintenance Management System). We need someone who can design, implement, and test the entire system, including data ingestion, feature engineering, model training, deployment, and monitoring. The ideal candidate will have a strong background in machine learning, deep learning, and cloud computing. We'll focus on minimizing false positives and maximizing prediction accuracy to minimize downtime and maintenance costs. A pilot project will be implemented on a subset of our machinery before full-scale deployment.