Musfiqur Rahman

PhD Candidate in Software Engineering at Concordia University & CREATE SE4AI Trainee

20250317_143649.jpg

I am a PhD Candidate in Software Engineering at Concordia University’s DAS Lab, supervised by Prof. Emad Shihab. My thesis, “Large Language Models in Coding: Generation, Detection, and Repair,” investigates how LLMs generate code, how to detect it, and how to repair it. This work includes SIEVE, a live contamination-aware corpus builder for LLM-generated code, and OpenClassGen, a large-scale dataset of LLM-generated Python classes published at EASE 2026. Before my PhD, I worked as an AI Engineer at Pentavere and as Lead Data Science Instructor at General Assembly, giving me hands-on experience applying machine learning in industry and teaching it to others.

When I am not coding (usually in Python) or writing, you might find me chasing after my three-year-old. You can also often spot me exploring the streets of Toronto and Montreal, camera in hand. Other times, I’m likely at a cozy café with my beautiful wife, taking a break from the joys and exhaustion of parenting, though we usually end up talking about our little one. If I’m alone, I might be at that same café, daydreaming about a pilgrimage to Mecca and Medina or a future culinary adventure in Japan with my family.

news

Jul 01, 2026 Actively seeking industry and research roles in ML/AI Engineering, Applied Science, and Data Science. Open to opportunities across Canada and the US.
Jun 10, 2026 Presented at the Doctoral Symposium at EASE 2026. Talk title: “From Observation to Explanation: Mechanistic Interpretability of LLM-Generated Code for Principled Repair.”
Apr 08, 2026 Paper accepted and published at EASE 2026 — OpenClassGen: A Large-Scale Open Dataset of LLM-Generated Python Classes. DOI: 10.1145/3816483.3816547.
Dec 17, 2025 Website is live!

latest posts

selected publications

  1. EASE
    OpenClassGen: A Large-Scale Open Dataset of LLM-Generated Python Classes
    Musfiqur Rahman, SayedHassan Khatoonabadi, and Emad Shihab
    In Proceedings of the 30th International Conference on Evaluation and Assessment in Software Engineering (EASE 2026), 2026
  2. arXiv
    Beyond Synthetic Benchmarks: Evaluating LLM Performance on Real-World Class-Level Code Generation
    Musfiqur Rahman, SayedHassan Khatoonabadi, and Emad Shihab
    arXiv preprint arXiv:2510.26130, 2025
  3. EASE
    The Impact of Environment Configurations on the Stability of AI-Enabled Systems
    Musfiqur Rahman, SayedHassan Khatoonabadi, Ahmad Abdellatif, and 2 more authors
    In Proceedings of the 29th International Conference on Evaluation and Assessment in Software Engineering, 2025