I need a full-cycle AI solution that continuously tracks asset health, flags early signs of degradation, and forecasts failures before downtime hits. My focus is AI-driven predictive maintenance across our equipment fleet, using the following data streams in real time (through Ignition from Inductive Automation): • Vibration sensors • Temperature sensors • Pressure sensors Existing time series data are stored on an SQL database and accessed through Ignition. Key goals 1. Start with a proof of concept on 1 process equipment and 10 tags (channels). 2. Build and train models that can spot mechanical and electrical failures as well as production-quality issues early. 3. Deploy the models in a lightweight service that runs 24/7, sends actionable alerts, and learns from new data to improve accuracy over time. 4. Provide a clear dashboard (web or desktop) that visualizes current condition, remaining useful life, and confidence levels for each monitored asset. I’m open to your preferred stack, but the workflow must integrate smoothly with common industrial protocols (e.g., OPC UA, Modbus) and be easy for my maintenance team to administer. Please include in your proposal: • A brief outline of your approach to data preprocessing, feature extraction, and model selection. • Examples of similar predictive-maintenance work you’ve delivered. • Estimated timeline for a functional prototype and a polished production release. Looking forward to collaborating on a robust, scalable system that keeps our operations running reliably.