Mohammad J Islam

Welcome to my academic and professional journey. I am currently a Ph.D. student at the University of California, Riverside supervised by Dr. Shaolei Ren where my research interest includes Online Optimization and Sustainable AI. Before this, I earned my MS degree in Computer Science from Western Michigan University in May 2021 where my research focused on secure systems design within Cyber-Physical Systems under the supervision of Dr. Shameek Bhattacharjee.

My professional experience includes working as a Software Engineer at Reve Systems Bangladesh from 2016 to 2019, where I gained valuable insights into software development and engineering practices. I completed my BS degree in Computer Science from Bangladesh University and Engineering and Technology in March 2016, laying the foundation for my career in technology.

On this platform, I share insights into my research interests, experiences, and projects in the fields of Online Optimization, Sustainable AI, and Secure Systems Design. Join me as I explore innovative solutions at the intersection of computer science and engineering.

Research Interest

Research Interest

Online & Robust Optimization.
Sustainable AI.
Cyber-Physical System.
Adversarial ML.

Education

Ph.D. in Computer Science, 2022-Present.

University of California Riverside, Riverside, CA, USA.

M.S. in Computer Science, 2019-2021.

Western Michigan University, Kalamazoo, MI, USA.

B.Sc. in Computer Science and Engineering, 2011-2016

Bangladesh University of Engineering and Technology, Dhaka Bangladesh.

Research Interest

Online & Robust Optimization.
Sustainable AI.
Cyber-Physical System.
Adversarial ML.

Education

Ph.D. in Computer Science, 2022-Present.

University of California Riverside, Riverside, CA, USA.

M.S. in Computer Science, 2019-2021.

Western Michigan University, Kalamazoo, MI, USA.

B.Sc. in Computer Science and Engineering, 2011-2016

Bangladesh University of Engineering and Technology, Dhaka Bangladesh.

Publication

Mohammad Jaminur Islam; Saminur Islam; Mahmud Hossain; Shahid Noor; S. M. Riazul Islam "Securing Blockchain Systems: A Layer-Oriented Survey of Threats, Vulnerability Taxonomy, and Detection Methods"
Pengfei Li; Mohammad Jaminur Islam; Shaolei Ren "A Case Study of Environmental Footprints for GenerativeAI Inference: Cloud versus Edge". (Accepted in ACM SIGMETRICS 2025 AI Crossroads (AI Crossroads 2025)
Mohammad Jaminur Islam; Richard T Meyer; Shameek Bhattacharjee "Poisoning Attacks against Quantile L1 Regression in CPS

Anomaly Detection Frameworks" (Accepted in IWSPA 25).

Mohammad Jaminur Islam; Shaolei Ren "Equity-Aware Spatial-Temporal Workload Shifting for Sustainable AI Data Centers"(Accepted in CCAI-NeurIPS 24).
Jianyi Yang; Pengfei Li; Mohammad Jaminur Islam; Shaolei Ren “Online Allocation with Replenishable Budgets: Worst Case and Beyond” Published in ACM Sigmetrics 2024.
Mohammad Jaminur Islam; Jose Paolo Talusan; Shameek Bhattacharjee; Francis Tiausas; Abhishek Dubey; Keiichi Yasumoto; Sajal K. Das “Scalable Pythagorean Mean based Incident Detection in Smart Transportation Systems” Published in ACM TCPS23.
Saminur Islam; Mohammad Jaminur Islam; Mahmud Hossain; Shahid Noor; Kyung-Sup Kwak; S. M. Riazul Islam “A Survey on Consensus Algorithms in Blockchain-Based Applications: Architecture, Taxonomy, and Operational Issues” Published in IEEE ACCESS23.
Mohammad Jaminur Islam; Jose Paolo Talusan; Shameek Bhattacharjee; Francis Tiausas; Sayyed Mohsen Vazirizade; Abhishek Dubey; Keiichi Yasumoto; Sajal K. Das “Anomaly-based Incident Detection in Large Scale Smart Transportation Systems” Published in IEEE/ACM ICCPS 2022. Best Paper Nominee.
Shameek Bhattacharjee; Mohammad Jaminur Islam; Sahar Abedzadeh “Robust Anomaly-based Attack Detection in Smart Grids under Data Poisoning Attacks” Published in ACM CPSS22

Work Experience

University of California, Riverside | Graduate Research Assistant | Graduate Teaching Assistant Fall2022-Present

LLM benchmarking-Cloud
- Pioneered a multi-dimensional benchmarking framework that not only assessed model performance in terms of latency but also quantified energy consumption and carbon emissions, providing a comparison of quantized and non-quantized LLMs like Gemma-2B and LLaMA-7B under real-world cloud conditions.
- Demonstrates that lower quantization such as 4bit or 8bit may not guarantee energy or carbon savings.
Context-Aware Workload Allocation in Datacenter
- Created an RNN model to predict future contexts for a data center such as workload, electricity price, water, and carbon efficiency using historical data for resource allocation.
Taught an Introduction to AI (CS170) course for 60 students, providing hands-on instruction in key algorithms and programming techniques, fostering a strong foundational understanding of artificial intelligence

Western Michigan University, Kalamazoo | Graduate Research Assistant | Graduate Teaching Assistant Fall2019- Sping 2021

Designed data-driven anomaly detection models for smart living things with added impact analysis of adversarial attacks on the system.

Reve Systems, Dhaka, Bangladesh | Software Engineer Aug 2016 - July 2019

Automation software development for e-services such as inventory and billing management.