Shoumik Saha

[CV]. [Google Scholar]. [LinkedIn].
Live and let live!

Shoumik_profile_pic.jpg

CS Ph.D. Student

UMD College Park

smksaha@umd.edu

I am a 3rd year Computer Science Ph.D. student at the University of Maryland - College Park, where I am fortunate to be advised by Prof. Soheil Feizi. Last summer, I have worked as a Scientist Intern at Amazon AWS. My research journey began with a focus on machine learning for security, particularly in malware detection. Over time, my interests have evolved toward security and reliability in machine learning. These days, I’m dedicated to enhancing the robustness and reliability of generative AI, especially large language models (LLMs).

If you check out my CV, you’ll see a consistent theme: I enjoy exploring challenges from both sides of the coin—attack and defense, red team and blue team, or however you’d like to frame it. Sounds interesting? Feel free to reach out to discuss my research or potential collaborations!

Before joining the Ph.D. program, I earned my B.Sc. from Bangladesh University of Engineering and Technology (BUET). I then gained valuable experience as a full-time lecturer at United International University and a part-time research assistant at BUET’s research lab.

news

Jan 19, 2025 Paper (co-authored) accepted to SaTML 2025!
Dec 20, 2024 Fall 2024: Completed all the coursework of Ph.D.

selected publications

  1. NEURIPS
    selfcheck_preview2-1.png
    LLM-Check: Investigating Detection of Hallucinations in Large Language Models
    Gaurang Sriramanan, Siddhant Bharti, Vinu Sankar Sadasivan, Shoumik Saha, Priyatham Kattakinda, and Soheil Feizi
    In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024
  2. ICML
    beast_preview2.png
    Fast Adversarial Attacks on Language Models In One GPU Minute
    Vinu Sankar Sadasivan, Shoumik Saha, Gaurang Sriramanan, Priyatham Kattakinda, Atoosa Chegini, and Soheil Feizi
    2024
  3. ICLR
    drsm_preview2-1.png
    DRSM: De-Randomized Smoothing on Malware Classifier Providing Certified Robustness
    Shoumik Saha, Wenxiao Wang, Yigitcan Kaya, Soheil Feizi, and Tudor Dumitras
    2023
  4. Computers & Security
    malign_preview2.png
    MAlign: Explainable static raw-byte based malware family classification using sequence alignment
    Shoumik Saha, Sadia Afroz, and Atif Hasan Rahman
    Computers & Security, 2024
  5. arxiv
    ml_malware_preview.png
    Demystifying Behavior-Based Malware Detection at Endpoints
    Yigitcan Kaya, Yizheng Chen, Shoumik Saha, Fabio Pierazzi, Lorenzo Cavallaro, David Wagner, and Tudor Dumitras
    2024