Abdullah Naveed

About me

I am a Ph.D. candidate in Computer Engineering at the University of Florida, advised by Dr. Guanpeng Li in the Dependable Systems Lab. My research lies at the intersection of dependable machine learning, high-performance computing, and scientific data systems.

I study how large language models and HPC applications behave under hardware and software faults, and I design low-overhead mechanisms that make these systems more reliable. At Argonne National Laboratory, I am advised by Dr. Sheng Di on hardware–software co-design for scientific data compression.

My current interests include:

Education

Experience

Full experience

Ongoing research

Dependable large language models

Characterizing soft-error behavior across LLM inference, parameter-efficient fine-tuning, and pretraining, while developing lightweight runtime detection.

Scientific data compression

Co-designing hardware and software compression methods for data-intensive scientific workflows.

Compiler-based HPC resilience

Developing LLVM-based silent-error detection using selective runtime monitoring for scientific and HPC applications.

Publications

All publications
  1. TRACE offline profiling and online monitoring workflowSC’26
    Low-Overhead Silent Error Detection for Large Language Model Inference at Scale

    A. Naveed, M. Sheikh, and G. Li

    [SC’26] Accepted; to appear

  2. Z-SPDK scientific data compression and storage workflowSC’26
    Z-SPDK: Accelerating Lossy Compressed I/O on NVMe-SSDs

    D. Ng, A. Naveed, D. Zhang, G. Li, S. Di, and X. Lu

    [SC’26] Accepted; to appear

  3. DebloatBench framework architectureSecDev’23
    Evaluating Container Debloaters

    M. Hassan, T. Tahir, M. Farrukh, A. Naveed, et al.

    [SecDev’23] IEEE Secure Development Conference

    PaperCode

Selected projects

All projects
  1. Multimodal fake-news classification architectureML
    Multimodal Fake News Classification

    Fuses BERT-based text features with CNN-derived visual representations for multimodal classification.

    Repository

  2. Compiler optimization project diagramLLVM
    Compiler Optimizations

    Implementations of research-driven LLVM optimization techniques for performance-focused workloads.

    Repository

  3. SOP violation detection multi-camera interfaceVision
    COVID SOP Violation Detection

    A real-time multi-camera system for mask detection, top-view projection, and density heatmaps.

    Repository