Work on multivariate time-series forecasting and anomaly detection for industrial energy data to power real-world monitoring and decision-support tools.

Scope and datasets:

• Energy consumption, PV production, and building temperature time series
• Industrial/commercial building context with real operational constraints

Models and methods:

• Machine Learning and Deep Learning for forecasting and anomaly detection
• Benchmarking multiple model families and evaluation protocols
• Optional LLM integration for insights/explanations

Expected deliverables:

• Robust forecasting model for energy signals
• Anomaly detection solution for operational monitoring
• Interactive dashboard (e.g., Streamlit/Dash) for analysis and reporting
• Research paper-quality write-up

Technical environment:

• Python, PyTorch, TensorFlow, scikit-learn
• Streamlit, Dash, Plotly, Gradio
• Git, Jupyter; cloud deployment optional

Profile:

• Final-year engineering student (AI, Data Science, CS, Energy, Automation)
• Strong Python and ML/DL skills (time series, classification, model evaluation)
• Dashboard development experience is a plus
• Good communication, teamwork, curiosity, autonomy

What we offer:

• Real-world Industry 4.0 and energy-tech projects
• Mentorship from an experienced AI/Data Science team
• Opportunity to co-author a scientific publication
• Immersion in a fast-growing scale-up; potential for future collaboration

How to apply:

• Email your CV/Resume and (optional) GitHub/Portfolio
• Important: mention the project title in the email subject

📧 Pour postuler: feres.jerbi@wattnow.io