Assistant Professor | AI Researcher | Industrial AI

Bridging foundational AI research with real-world industrial deployment to drive the next generation of intelligent manufacturing and predictive maintenance systems. I lead research at the intersection of Industrial AI, Generative Models, and AI Agents, transforming how industries monitor, diagnose, and maintain critical machinery.

Assistant Professor (RTD-A) at Politecnico di Torino – Department of Mechanical and Aerospace Engineering (DIMEAS)
Research Group – Industrial Systems Engineering and Design (ISED)

Leading the development of production-ready AI systems for condition monitoring, fault diagnosis, and anomaly detection in industrial rotating machinery. My work spans the full AI lifecycle: from experimental data acquisition to deployed edge solutions.

Pioneering the application of cycleGANs and generative models to create high-fidelity synthetic data for damaged machinery in 2023, dramatically reducing the need for costly experimental campaigns while maintaining diagnostic accuracy.

Impact:

• Demonstrate zero-shot learning for bearing fault diagnosis using synthetic data
• Reduced data acquisition costs by >70% through generative approaches

Building next-generation diagnostic systems leveraging Large Language Models, Transfer Learning, and Few-Shot Learning for industrial applications. Creator of PoC AI agents using the Model Context Protocol (MCP).

Innovation Highlights:

• 🔥 predictive-maintenance-mcp: Open-source AI agent integrating vibration analysis, datasets, and diagnostic workflows
• 📱 Featured on Medium with international outreach
• 🎓 Teaching PhD-level courses on advanced AI techniques for machinery diagnosis (UPC Barcelona, July 2025)

• Core ML/DL Stack: PyTorch | TensorFlow | Scikit-learn
• Deep Learning Architectures: CNNs | GANs | cycleGANs | Transfer Learning
• Advanced Methods: Anomaly Detection | Few-Shot Learning
• Explainable AI: SHAP | LIME | GradCAM

• Large Language Models: Claude | GPT-4 | Gemini
• MCP Development: Model Context Protocol implementation and server design
• Prompt Engineering: Advanced prompting strategies and guardrails design
• AI Orchestration: AI Agent deployment and workflow automation

• Research Focus: Developing foundation models specifically designed for machine diagnosis and maintenance
• Transfer Learning: Adapting pre-trained models to industrial diagnostic tasks
• Domain Adaptation: Building robust models that generalize across machinery types

• Edge AI: working on STM32 | STWIN.BOX | Embedded ML systems
• Cloud Infrastructure: working on AWS deployment and scaling
• DevOps & MLOps: Docker | Git | CI/CD pipelines thorugh coding assistant
• Signal Processing: Time-series analysis | Vibration monitoring | Feature extraction | Advanced time-frequency modeling

• FEM/FEA: Abaqus | ANSYS Mechanical
• CFD & FSI: ANSYS Fluent | Fluid-Structure Interaction coupling
• Condition Monitoring: Vibration analysis | Modal analysis
• Diagnostics: Motor Current Signature Analysis (MCSA) | Bearing fault detection
• Fatigue Analysis: nCode DesignLife | FE-Safe

🌟 Zero-Shot Generative AI for Rotating Machinery (2023)
Applied Sciences, MDPI First demonstration of cycleGANs generating synthetic bearing fault data from simple simulations

🧠 Transfer Learning from Audio CNNs to Industrial Diagnostics (2023)
Sensors, MDPI
Repurposing pre-trained sound classification networks for bearing fault detection – 95%+ accuracy

🔍 Explainable AI for Industrial Condition Monitoring (2023)
Applied Sciences, MDPI | 🏆 Best Paper Award 2025
SHAP-based feature importance analysis for interpretable SVM/kNN fault classifiers

⚙️ Large-Scale Industrial Bearing Test Rig Design (2022)
Machines, MDPI
Self-balancing layout enabling controlled fault injection in medium-sized bearings

🤖 Toward LLM-Based AI Agents for Predictive Maintenance (2025)
Model Context Protocol integration with diagnostic workflows

📊 Open Datasets for Medium-Sized Industrial Bearings (2024, 2025)
Benchmark datasets enabling reproducible research in bearing diagnostics

📈 Novelty Detection for Industrial Systems (2025)
Unsupervised/semi-supervised approaches for anomaly detection in machinery

• Co-supervisor: 1 Ph.D. student
• Thesis supervision: MSc + BSc
• Research grants: Scientific supervisor for AI in autonomous LLM-based diagnostic systems

• Advanced AI for Machinery Diagnosis (PhD level, UPC Barcelona)

• 📝 Guest Editor for Special Issues in Applied Sciences and Electronics (MDPI)
• 🔬 Reviewer for 15+ international journals (IEEE, ASME, Elsevier, MDPI)

1. AI Agents for Industrial Maintenance 🤖
Developing production-ready LLM-based agents that integrate diagnostic tools, knowledge bases, and real-time monitoring systems through MCP architecture.

2. Synthetic Data Ecosystems 🎨
Building comprehensive pipelines for generating, validating, and deploying synthetic fault data across multiple machinery types.

3. Explainable Anomaly Detection 🔍
Advancing unsupervised learning methods with built-in interpretability for root cause analysis in industrial settings.

4. Edge Intelligence for Rotating Systems 📱
Deploying lightweight ML models on embedded systems (STM32, STWIN.BOX) for real-time condition monitoring.

5. Digital Twins & Physics-Informed ML 🔧
Combining FEA/CFD simulations with deep learning for hybrid modeling of bearing systems and structural components.

predictive-maintenance-mcp
MCP-based AI agent for predictive maintenance workflows. Integrates vibration analysis tools, bearing fault datasets, and diagnostic documentation.
Stack: Python, Claude AI, Model Context Protocol

Medium Articles
Technical deep-dives on AI agents, MCP architecture, and industrial AI deployment strategies.

I'm actively seeking collaborations in:

✅ Research Partnerships: Joint projects on AI for industrial systems, digital twins, generative models
✅ Industrial Consulting: Predictive maintenance strategy, AI deployment, technology assessment
✅ Student Supervision: MSc/PhD thesis co-supervision in Industrial AI, ML for diagnostics
✅ Open Source: Contributing to industrial AI frameworks, dataset curation, benchmarking

• 🏢 Industrial partners interested in piloting next-gen AI diagnostic systems
• 🎓 Research collaborations on LLMs for technical domains and synthetic data generation
• 💡 Funding opportunities (EU programs, national grants, public-private consortia)
• 🚀 Entrepreneurial ventures in deep-tech AI for manufacturing and Industry 4.0

Email: luigi.dimaggio@polito.it
Institution: Politecnico di Torino
Location: Turin, Italy | Open to hybrid/remote collaboration