
DebloatBench
Built a benchmarking framework for container debloaters (SPEAKER, CONFINE, SlimToolkit) that measures system-call reduction, CVE mitigation, correctness, and image size.
I’m a Ph.D. student in Computer Engineering at the University of Florida. My research centers on fault tolerance and error resilience for Large Language Models (LLMs) and high-performance computing (HPC) applications, with a complementary focus on data compression for scientific workloads.
I currently collaborate with Argonne National Laboratory on Transformer-based lossy compressors for Advanced Photon Source (APS) X-ray data, working with Dr. Sheng Di and Prof. Guanpeng Li. Previously, I completed my MCS at The University of Iowa (2025) and earned a BS in Computer Science with Distinction from LUMS, Pakistan.
Interests: machine learning systems, reliability for safety-critical AI, scientific data compression, and LLVM-based tooling for dependable systems.
email: abdullahnaveed102@gmail.com
location: Gainesville, FL, USA
University of Florida, Gainesville, FL, USA
Advisor: Prof. Guanpeng Li
Research Lab: Dependable Systems Lab
Research Areas: Fault Tolerance, Error Resilience for Large Language Models (LLMs), HPC Dependability, Scientific Data Compression
The University of Iowa, Iowa City, IA, USA
GPA: 3.90 / 4.00
Selected Coursework: Applied Machine Learning, Compiler Optimization for HPC, Distributed Algorithms, Functional Programming & Algorithm Design, Data Visualization & Technologies
Lahore University of Management Sciences (LUMS), Lahore, Pakistan
GPA: 3.61 / 4.00
Graduated with Distinction
Selected Coursework: Machine Learning, Computer Vision, Data Mining, Data Science, Network-Centric Computing, Blockchain Technology
Built a benchmarking framework for container debloaters (SPEAKER, CONFINE, SlimToolkit) that measures system-call reduction, CVE mitigation, correctness, and image size.
Collaborating with Dr. Sheng Di and Prof. Guanpeng Li on transformer-based lossy compression for APS X-ray data; targeting high-throughput performance while minimizing scientific distortion.
Jun 2025 – Present
Fault tolerance & error resilience for LLMs in safety-critical settings; designing lightweight monitoring and error-detection mechanisms to improve reliability.
Aug 2025 – Present
Built LLVM-based tools for instruction-level resilience analysis; evaluated compiler/ runtime techniques for fault detection in HPC workloads (dependable systems).
Aug 2023 – Aug 2025
Mentored 30+ students on MySQL projects, query debugging, and relational schema design through weekly labs and office hours.
Aug 2023 – Dec 2023
Guided 100+ students via tutorials and office hours; clarified programming assignments and reinforced algorithmic problem-solving skills.
Sep 2022 – Jan 2023