Ehsan Yousefzadeh-Asl-Miandoab

Office: 4E04

Rued Langgaards Vej 7

Copenhagen, 2300

I am a computer systems researcher interested in efficient computing, AI infrastructure, resource management, and system performance. My research focuses on understanding, designing, and improving modern computing systems, with an emphasis on efficiency, performance, and dependability.

My recent work focuses on GPU resource management and efficient AI infrastructure, where I develop measurement-driven methods, predictive models, and runtime systems to improve resource utilization and system efficiency.

I am currently a postdoctoral researcher in the ⁠Resource-Aware Data Systems (RAD) Group at the ⁠IT University of Copenhagen.

I am currently exploring faculty, postdoctoral, research scientist, and industry R&D opportunities starting in 2027, where my background in computer systems and efficient computing can contribute to new research and engineering challenges.

news

Oct 06, 2026	Our AEGIS poster has been accepted for presentation at D3A 4.0! Looking forward to presenting our work and connecting with the Danish AI and computer systems community.
Jun 26, 2026	The DOI is now available for our DBTest 2026 paper, Benchmarking Column Statistics for Analytical Query Pruning.

latest posts

Aug 02, 2026	Binary Search Beyond Sorted Arrays: Searching the Answer Space
Aug 18, 2025	Computer Graphics Story for a Better Understanding of Modern GPUs
Aug 13, 2025	SLURM and Kubernetes: A Beginner’s Guide to Resource Management Systems
Mar 25, 2024	NVIDIA GPUs story
Oct 14, 2023	“A Study of Checkpointing in Large Scale Training of Deep Neural Networks” paper summary

selected publications

EuroMLSys

GPU Memory and Utilization Estimation for Training-Aware Resource Management: Opportunities and Limitations

Ehsan Yousefzadeh-Asl-Miandoab, Reza Karimzadeh, Danyal Yorulmaz, and 2 more authors

In Proceedings of the Sixth European Workshop on Machine Learning and Systems, 2026

DOI Bib Code Dataset

@inproceedings{yousefzadeh2026gpumemutil,
  title = {GPU Memory and Utilization Estimation for Training-Aware Resource Management: Opportunities and Limitations},
  author = {Yousefzadeh-Asl-Miandoab, Ehsan and Karimzadeh, Reza and Yorulmaz, Danyal and Ibragimov, Bulat and T{\"o}z{\"u}n, P{\i}nar},
  booktitle = {Proceedings of the Sixth European Workshop on Machine Learning and Systems},
  pages = {127--138},
  year = {2026},
  doi = {10.1145/3805621.3807621},
  dataset = {https://huggingface.co/datasets/ehyo/GPU-Resources-Estimation-for-Deep-Learning-Training-Tasks},
}

EuroMLSys

Profiling and Monitoring Deep Learning Training Tasks

Ehsan Yousefzadeh-Asl-Miandoab, Ties Robroek, and Pınar Tözün

In Proceedings of the Third Workshop on Machine Learning and Systems, 2023

DOI Bib Code

@inproceedings{yousefzadeh2023profiling,
  title = {Profiling and Monitoring Deep Learning Training Tasks},
  author = {Yousefzadeh-Asl-Miandoab, Ehsan and Robroek, Ties and T{\"o}z{\"u}n, P{\i}nar},
  booktitle = {Proceedings of the Third Workshop on Machine Learning and Systems},
  pages = {18--25},
  year = {2023},
  doi = {10.1145/3578356.3592589},
}

TPDS

OSM: Off-Chip Shared Memory for GPUs

Sina Darabi, Ehsan Yousefzadeh-Asl-Miandoab, Negar Akbarzadeh, and 4 more authors

IEEE Transactions on Parallel and Distributed Systems, 2022

DOI Bib

@article{darabi2022osm,
  title = {OSM: Off-Chip Shared Memory for GPUs},
  author = {Darabi, Sina and Yousefzadeh-Asl-Miandoab, Ehsan and Akbarzadeh, Negar and Falahati, Hajar and Lotfi-Kamran, Pejman and Sadrosadati, Mohammad and Sarbazi-Azad, Hamid},
  journal = {IEEE Transactions on Parallel and Distributed Systems},
  volume = {33},
  number = {12},
  pages = {3415--3429},
  year = {2022},
  doi = {10.1109/TPDS.2022.3154315},
}