Hadeel Albahar

Assistant Professor, Department of Computer Engineering, Kuwait University and Visiting Researcher, Department of Computer Science, Virginia Tech

Systems for ML GPU Scheduling Tensor Caching LLM Inference HPC Storage Systems Privacy-Preserving ML

I build systems for large-scale machine learning: GPU scheduling, tensor caching, and storage efficiency for deep learning training. My ongoing work focuses on the memory wall in LLM training and inference.

Hadeel Albahar

About

I am an Assistant Professor in the Department of Computer Engineering at Kuwait University.

My research designs and optimizes systems for large-scale machine learning workloads including GPU scheduling, tensor caching, and storage efficiency for deep learning training. I am broadly interested in the intersection of High Performance Computing, Systems for ML, and ML for Systems.

My ongoing work explores how to improve LLM training and inference in the face of the memory wall, as well as problems at the intersection of systems and privacy-preserving machine learning.

I am currently a Visiting Researcher at Virginia Tech, where I previously completed my Ph.D. in Computer Engineering advised by Dr. Ali R. Butt in the Distributed Systems & Storage Laboratory.

Education

  • Ph.D., Computer Engineering Virginia Tech
  • M.S., Computer Engineering Columbia University
  • B.S., Computer Engineering Kuwait University · Class Honors

Research

Systems for ML · GPU Clusters

Heterogeneity-Aware GPU Scheduling

Schedulers that exploit hardware diversity in multi-GPU clusters to minimize DL training time. Combines learned predictors with optimization-based dispatch.

GPU SchedulingDL WorkloadsHeterogeneous Clusters

LLM Systems · Memory Wall

Tensor Caching for LLM Training

Addressing memory bottlenecks in large-model training and inference through heterogeneous resource-aware tensor caching across GPU HBM, CPU DRAM, and NVMe.

Tensor CachingLLM TrainingMemory Hierarchy

Storage · Container Systems

Cloud & Container Storage Systems

High-performance deduplication for container registries and Docker storage, and LSM-tree stores that integrate local storage with cloud storage for fast, efficient data access.

DeduplicationDockerCloud Storage

Emerging · Privacy × Systems

Systems for Privacy-Preserving ML

Efficient infrastructure for federated and confidential training workloads at scale, at the intersection of systems design and privacy-preserving machine learning.

Privacy-Preserving MLFederated Learning

Publications

Peer-reviewed publications, reverse chronological. Full list on Google Scholar and ORCID.

Conference Papers

  1. 7
    SoCC '25 Hybrid Learning and Optimization-Based Dynamic Scheduling for DL Workloads on Heterogeneous GPU Clusters

    Shruti Dongare, Redwan Ibne Seraj Khan, Hadeel Albahar, Nannan Zhao, Diego Meléndez-Maita, Ali R. Butt. ACM Symposium on Cloud Computing, 2025.

  2. 6
    SoCC '25 10Cache: Heterogeneous Resource Aware Tensor Caching and Migration for LLM Training

    Sabiha Afroz, Redwan Ibne Seraj Khan, Hadeel Albahar, Jingoo Han, Ali R. Butt. ACM Symposium on Cloud Computing, 2025.

  3. 5
    CCGrid '22 SchedTune: A Heterogeneity-Aware GPU Scheduler for Deep Learning

    Hadeel Albahar, Shruti Dongare, Yanlin Du, Nannan Zhao, Arnab K. Paul, Ali R. Butt. IEEE/ACM CCGrid, 2022.

  4. 4
    Cluster '21 Building A Fast and Efficient LSM-tree Store by Integrating Local Storage with Cloud Storage

    Peng Xu, Nannan Zhao, Jiguang Wan, Wei Liu, Shuning Chen, Yuanhui Zhou, Hadeel Albahar, et al. IEEE Cluster, 2021.

  5. 3
    ATC '20 DupHunter: Flexible High-Performance Deduplication for Docker Registries

    Nannan Zhao, Hadeel Albahar, Subil Abraham, Keren Chen, Vasily Tarasov, Dimitrios Skourtis, Lukas Rupprecht, Ali Anwar, Ali R. Butt. USENIX ATC, 2020.

  6. 2
    Cluster '19 Large-Scale Analysis of the Docker Hub Dataset

    Nannan Zhao, Vasily Tarasov, Hadeel Albahar, Ali Anwar, Lukas Rupprecht, Dimitrios Skourtis, Amit S. Warke, Mohamed Mohamed, Ali R. Butt. IEEE Cluster, 2019.

  7. 1
    NER '17 In-vivo transcranial ultrasound imaging of induced Substantia Nigra hyperechogenicity using adaptive sparse Third Order Volterra Filter

    Mohamed Almekkawy, James Cunningham, Yi Song, Hadeel Albahar, Thyagarajan Subramanian. IEEE NER, 2017.

Journal Articles

  1. 4
    TOS '24 An End-to-End High-Performance Deduplication Scheme for Docker Registries and Docker Container Storage Systems

    Nannan Zhao, Muhui Lin, Hadeel Albahar, Arnab K. Paul, et al. ACM Transactions on Storage, 2024.

  2. 3
    TACO '22 Building a fast and efficient LSM-tree store by integrating local storage with cloud storage

    Peng Xu, Nannan Zhao, Jiguang Wan, Wei Liu, et al. ACM TACO, 2022.

  3. 2
    TPDS '20 Large-Scale Analysis of Docker Images and Performance Implications for Container Storage Systems

    Nannan Zhao, Vasily Tarasov, Hadeel Albahar, Ali Anwar, Lukas Rupprecht, et al. IEEE TPDS, 2020.

  4. 1
    DII '18 High intensity focused ultrasound: The fundamentals, clinical applications and research trends

    IA Shehata Elhelf, H Albahar, U Shah, A Oto, E Cressman, M Almekkawy. Diagnostic and Interventional Imaging, 2018.


Teaching

CpE 363

Introduction to Embedded Systems

Fall'23 · Summer'24 · Spring'25 · Summer'25 · Fall'26

CpE 445

Operating System Principles

Fall'23 · Spring'24

ENGR 310

Engineering Ethics

Spring'24 · Fall'24 · Fall'26

CpE 262

Fundamentals of Digital Logic

Fall'24


Community & service

A selection of recent roles; full list in my CV.

Program Committees

  • SoCC '26ACM Symposium on Cloud Computing
  • HiPC '25IEEE Int'l Conf. on High Performance Computing
  • Cluster '25IEEE Int'l Conf. on Cluster Computing
  • IPDPS '25IEEE Parallel & Distributed Processing Symposium
  • REX-IOExtreme-Scale I/O Workshop ('24–'26)

Leadership & Outreach

  • 2026–Co-Founder, Women in HPC MENA
  • 2025/26Mentor, VA-WHPC Student Lightning Talks
  • 2025Coordinator, Seminar Series, Computer Engineering, KU
  • 2024/25Volunteer Instructor, Linux & Docker, KARS, KU
  • 2022Virtual Hopper Volunteer, Grace Hopper Celebration

Contact

hadeel.albahar AT ku.edu.kw
South Engineering Building, 2nd floor, D1 area, room 160
Sabah Al-Salem University City, Kuwait