Grants
Research funding and grants.
- Sep 2026
Huawei Research Grant on HarmonyOS
Funding for four PhD positions starting in September 2026, with EUR 1.04M over two years and the possibility of extension for another two years.
- Jun 2026
Funding for one PhD position starting in June 2026, with total funding of EUR 350K.
In this subproject, we study agentic mid-training methods for foundation models, with applications to web agents and operating-system agents. The project connects to the broader Bavarian AI Foundation Model Initiative, which develops open multimodal foundation models for science, public administration, and society.
- Jun 2025
Huawei Internal Funding: Agentic Memory
Funding for one PhD student over three years, with total funding of EUR 400K.
This project investigates agentic memory for multimodal agents, with a focus on multimodal memory for long-form video understanding. The goal is to design agents that can construct, update, retrieve, and verify memories over long temporal horizons, enabling more robust long-context reasoning.
- Aug 2026
Siemens Research Grant: Multi-Agentic Systems for Industrial Production Applications
Funding for two PhD positions, one at LMU and one at Siemens, with total funding of EUR 600K.
The project studies agentic system optimization and industrial multimodal agentic systems for production applications. It focuses on how multiple specialized agents can perceive industrial states, coordinate decisions, and optimize planning and execution in realistic production workflows.
- Apr 2026
Funding for 20 PhD positions over five years, with the potential to extend to ten years. Our group will have three PhD positions over the full ten-year period, corresponding to EUR 1M in funding, in collaboration with Prof. Dr. Eyke Hüllermeier.
The subproject focuses on memory-augmented vision-language-action models, especially memory during embodied control, including episodic and procedural memory for VLA models.
- Jun 2025
Funding for one PhD position over four years, with total funding of EUR 400K.
This project studies memory-augmented LLMs for embodied control, with emphasis on continual learning, external memory, and the reuse of past interactions, task plans, and corrections. The broader goal is to enable embodied agents to act more robustly in long-horizon, changing environments.
- Jun 2024
Huawei Internal Funding: LLM Post-Training
Funding for one PhD student over three years, co-supervised with Dr. Xun Xiao and Prof. Dr. Volker Tresp.
This project focuses on post-training methods for large language models, including data-efficient adaptation, reasoning-oriented training, and agent-oriented model improvement.
- Jul 2022
QCHALLenge: Quantum-Classical Hybrid Optimization Algorithms for Logistics and Production Line Management
Funding for half a PhD position over three years, with EUR 150K in funding, in collaboration with Prof. Dr. Volker Tresp.
The subproject focuses on neural architecture search for variational quantum circuits and applications to industrial logistics and combinatorial optimization problems. More broadly, QCHALLenge develops hybrid quantum-classical algorithms, concepts, and tools for production and logistics use cases.
- Apr 2021
Funding for one PhD position at LMU, with EUR 350K in funding, in collaboration with Prof. Dr. Volker Tresp.
The project investigates how reinforcement learning can be implemented on quantum computers for industrial applications, including process control, distributed automation in smart factories, and production planning. It also develops algorithms, benchmarks, and libraries to make quantum reinforcement learning usable in industrial settings.
