Master’s thesis “Energy-efficient IoT–edge–cloud continuum infrastructures for dynamic AI service support”

LLMs are increasingly deployed not just in centralized cloud data centres but across a mobile and heterogeneous edge-cloud continuum. Their deployment is driven by requirements on latency, privacy, and cost (OpEx, CapEx). Yet their energy footprint across this distributed infrastructure remains poorly understood and largely unmanaged.

This master thesis addresses energy-aware scheduling of LLM inference workloads across a heterogeneous edge-cloud continuum. The work is developed within the fortiss Industrial IoT Lab, where an energy observability infrastructure is already in place. The thesis extends this infrastructure and contributes to bridging LLM-level energy characterisation with infrastructure-level energy control, through placement, routing, and scheduling decisions that jointly optimise inference quality, latency, and energy efficiency.

Your tasks:

Familiarize yourself with LLM inference (tiny, small LLMs), edge-cloud continuum architectures, and energy-aware resource management.
Extend the existing energy observability infrastructure in the fortiss IIoT Lab with energy models (Kepler, Prometheus) for LLM inference workloads.
Integrate the energy observability layer with an SDN control plane to enable energy-aware traffic engineering., and into the network digital twin available in the lab.
Design and implement LLM inference scheduling strategies that jointly optimise inference quality, latency, and energy efficiency across hardware configurations and quantization strategies.
Evaluate the implemented framework against relevant benchmarks and document the design, contributing to open tools and datasets.

Your profile:

Master student in Computer Science, Electrical Engineering, Communications Engineering, or a related field.
Background in machine learning, distributed systems, and/or computer networking (knowledge of LLMs, edge computing, or cloud infrastructure is a plus).
Practical programming experience in Python, Java, C/C++, or similar.
Familiarity with ML frameworks (e.g., PyTorch, HuggingFace) and/or network emulation tools is desirable.
Strong analytical skills and interest in sustainable AI, systems optimisation, and energy efficiency.
Excellent communication skills in English; German is a plus.

Our offer:

An exciting and innovative open research environment within a nationally funded R&D project.
Direct collaboration with the Swiss partner ZHAW and academic/industry research groups in a bilateral CH-DE project.
International, dynamic work environment with highly qualified and motivated colleagues.
Opportunity to publish research results and contribute to open tools and datasets for sustainable AI infrastructure.
Note: this master thesis position is not remunerated.

Have we piqued your interest?

If so, please submit your application, including a cover letter, a detailed resume, and a recent transcript.

Job ID: IIoT-MSC-02-2026
Contact: Prof. Dr. Rute Sofia personal@fortiss.org