Sadman Jashim Sakib

Research Interests

My work lies at the intersection of empirical software engineering, program analysis, and AI-driven automation. I focus on building practical tools and novel approaches to make complex software systems more reliable, maintainable, and easier to evolve.

Key areas include:

• Software Dependency Management
  Resolving version conflicts, improving ecosystem health, and reducing breakage in large-scale Python and Java (Maven) projects. My M.Sc. thesis introduced SMT-based techniques (Satisfiability Modulo Theories) to efficiently solve dependency resolution problems.

• Program Analysis and AST-based Techniques
  Leveraging abstract syntax trees (ASTs) and static/dynamic analysis to understand and transform code behavior at scale.

• Large Language Models for Software Engineering
  Exploring LLMs for automated debugging, program repair, code recommendation, and intelligent developer assistance.

• Automated Debugging and Program Repair
  Developing AI-powered methods to detect, diagnose, and fix bugs with minimal human intervention.

• Reproducibility of Python Software Systems
  Tackling environment inconsistencies, dependency drift, and build failures to ensure reproducible scientific and production workflows.

A signature part of my recent work is tooool — an open-source prototype that transforms real-world software engineering challenges (dependency solving, configuration problems, etc.) into Boolean Satisfiability (SAT) problems solvable by high-performance SMT solvers. This bridges classical constraint solving with modern SE pain points.

Recent Work & Publications

• Toward Better Dependency Management in Python Projects (M.Sc. Thesis, University of Windsor, 2025)
  Introduced SMTpip — an SMT-solver-based tool for conflict-free dependency resolution in Python.

• Understanding the Popularity of Packages in Maven Ecosystem (MSR 2025)
  Large-scale empirical study (103k+ packages) analyzing factors driving package adoption in Java/Maven.

I’m passionate about turning hard SE problems into solvable mathematical formulations and building tools that developers actually use. Currently based in Toronto, Ontario — open to PhD positions, collaborations, or discussions in automated SE, SE4AI, or dependency ecosystems.

Reach out via LinkedIn, GitHub, or email if our interests align!