Public summary
Contribute to innovative research combining Large Language Models and Time-Series Foundation Models to improve forecasting and decision-making in Home Energy Management Systems. This master thesis involves integrating a time-series model with an LLM ecosystem to enable interactive what-if analysis, evaluating system parameters and model performance in real scenarios using cutting-edge AI techniques. The role requires strong programming skills in Python and knowledge of machine learning, particularly time-series forecasting.
Location and work setup
- Location
- Renningen
- Remote status
- Hybrid
- German requirement signal
- No German Required Detected
- Detected job language
- English
Responsibilities
- Integrate a specialized time-series foundation model with an existing language model agent ecosystem using Model Context Protocol (MCP). - Abstract numerical time-series predictions into semantic forms for effective reasoning by LLM agents. - Design an interactive evaluation loop for exploring system parameters and assessing outcomes. - Systematically evaluate scenarios and data representations to measure benefits of specialized models over pure LLM reasoning. - Conduct independent scientific research and analysis throughout the thesis project.
Qualifications
- Enrollment in a master's program in Computer Science, Data Science, Software Engineering, Electrical Engineering, Mathematics, or related fields. - Strong proficiency in Python, especially with data and time-series libraries such as Pandas and NumPy. - Familiarity with asynchronous programming in Python (e.g., asyncio). - Solid understanding of AI and machine learning, with focus on time-series models and foundation models. - Knowledge or experience with agent-based workflows. - Highly self-motivated with strong analytical skills for independent research. - Excellent English language skills.