Tanja Harbaum

About

Tanja Harbaum is a researcher at the forefront of efficient deep learning deployment on embedded and distributed systems. Her primary research areas include convolutional neural network (CNN) inference partitioning, automated workload distribution, and resource-constrained embedded computing. Harbaum’s major contribution is the development of CNNParted, an open-source framework that enables efficient partitioning of CNN inference across multiple compute nodes in distributed embedded systems—critical for applications like robotics and autonomous driving where flexibility and robustness are paramount. Her work addresses the challenge of deploying data-flow-centric deep neural networks (DNNs) on resource-limited hardware by automating the partitioning process, as demonstrated in her 2024 paper on automated DNN inference partitioning. With her most-cited work accumulating 11 citations and her follow-up study garnering 6, Harbaum’s research is gaining traction in the embedded AI community. Her notable achievement lies in bridging the gap between high-performance deep learning and practical, real-time embedded systems, offering tangible solutions for scalable, distributed AI inference.

Research Focus

Key Achievements

Top Papers

1. 1
   CNNParted: An open source framework for efficient Convolutional Neural Network inference partitioning in embedded systems

   11 citations · 2023
2. 2
   Automated Deep Neural Network Inference Partitioning for Distributed Embedded Systems

   6 citations · 2024

Key Collaborators

Contact & Links