Postdoc: Edge-Cloud Foundation Model Inference

Communication, Control and Automation; Aarhus; On-site; DKK 420,000 - 480,000; Full time

Job summary

Aarhus University’s Department of Electrical and Computer Engineering invites applications for a 2-year postdoctoral position in adaptive and distributed foundation model inference across the computing continuum. The role focuses on resource-aware, edge-to-cloud inference and scalable deployment of large foundation models.

You will work with Prof. Qi Zhang on MoE placement, token compression, and distributed inference strategies, contributing to cutting-edge research and publishing in top venues.

Qualifications

• PhD in computer science, electrical engineering, or related field.
• Strong background in deep learning and transformers.
• Proficient in Python and PyTorch.
• Experience with distributed systems and edge computing.
• Strong publication record and excellent English communication.
• Experience with stream data or collaborative research environments is a plus.

Responsibilities

• Contribute to distributed foundation model inference research.
• Work on Mixture-of-Experts placement and scheduling for inference acceleration.
• Investigate token or feature compression for efficiency.
• Advance resource-aware and communication-efficient inference.

Skills

PhD in Computer Science

Deep Learning

Python

PyTorch

Distributed systems

Publications

English proficiency

Education

PhD in Computer Science / Electrical Engineering / related field

Tools

PyTorch

Job description

Aarhus University’s Department of Electrical and Computer Engineering invites applications for a 2-year postdoctoral position in adaptive and distributed foundation model inference across the computing continuum. The role focuses on resource-aware, edge-to-cloud inference and scalable deployment of large foundation models.

You will work with Prof. Qi Zhang on MoE placement, token compression, and distributed inference strategies, contributing to cutting-edge research and publishing in top venues.