Postdoc: Edge–Cloud Foundation Model Inference

Aarhus University; Aarhus; On-site; DKK 420,000 - 520,000; Full time

Job summary

Aarhus University invites applications for a 2-year postdoctoral position in Adaptive and Distributed Foundation Model Inference across the Computing Continuum. The candidate will advance resource-efficient, distributed inference leveraging edge and cloud resources, working within a dynamic research group at ECE AU.

The role focuses on distributed inference, MoE placement, and inference optimization, with opportunities to publish in top venues and collaborate with academic and industry partners

Qualifications

• PhD in Computer Science, Computer Engineering, Electrical Engineering or a related field.
• Strong background in Deep Learning (e.g., Transformers, foundation models).
• Strong programming skills in Python and experience with deep learning frameworks (e.g., PyTorch).
• Experience with distributed systems and edge AI.
• Strong publication record in reputable conferences or journals relative to career stage.
• Excellent written and spoken English communication skills.

Responsibilities

• Contribute to distributed foundation model inference research across the computing continuum.
• Develop and evaluate distributed inference strategies and MoE placement for acceleration.
• Investigate token compression and feature compression for efficient inference.
• Publish results and collaborate with partners within Horizon Europe project.

Skills

PhD in Computer Science/Engineering

Deep learning (Transformers)

Python

PyTorch

Distributed systems

English communication

Education

PhD in CS/CE/EE or related

Tools

PyTorch

Python

Job description

Aarhus University invites applications for a 2-year postdoctoral position in Adaptive and Distributed Foundation Model Inference across the Computing Continuum. The candidate will advance resource-efficient, distributed inference leveraging edge and cloud resources, working within a dynamic research group at ECE AU.

The role focuses on distributed inference, MoE placement, and inference optimization, with opportunities to publish in top venues and collaborate with academic and industry partners