Brahim Mahmoudi
Portrait of Brahim Mahmoudi

PhD Student in Software Engineering

Brahim MAHMOUDI

Passionate about artificial intelligence and software quality, I work on making AI-based systems more reliable, maintainable and robust.

  • École de technologie supérieure (ÉTS)
  • Montréal, Canada
Explore my research Get in touch
Brahim Mahmoudi in a rugby jersey in front of Saint Joseph's Oratory in Montréal, at night
Provincial & national champion 2024ÉTS Piranhas rugby team

About

Researcher, engineer & rugby player

My name is Brahim Mahmoudi, and I am currently a PhD student in Software Engineering at the École de technologie supérieure (ÉTS) in Montréal.

Passionate about artificial intelligence and software quality, my research focuses on improving the reliability, maintainability, and robustness of AI-based systems. Specifically, I work on the detection of AI-specific code smells through static code analysis and the design of a domain-specific language (DSL) that helps identify and prevent common pitfalls in AI codebases.

I hold an engineering degree in computer science from INSA Lyon, and I continued my academic path with a research-based master’s in software engineering at ÉTS, where I specialized in software quality, program analysis, and the modernisation of intelligent systems.

Outside of research, I am also a student-athlete, proudly representing the ÉTS rugby team, the Piranhas, with whom I became both provincial and national champion in 2024. This dual commitment has shaped my values around discipline, perseverance, and teamwork, which I bring both to the lab and to the field.

  • Now PhD, Software Engineering ÉTS, Montréal
  • Master’s Research M.Sc., Software Engineering ÉTS, Montréal
  • Engineering Engineering degree, Computer Science INSA Lyon, France
  • AI code smells
  • Static analysis
  • Domain-specific languages
  • LLM-based systems
  • Software quality
  • Microservices

PhD Project

Research

Specifying and detecting the pitfalls that make AI-based and LLM-based software fragile.

Overview diagram of the SpecDetect4AI approach

arXiv preprint 2025

SpecDetect4AI

Automated Detection of AI Code Smells Using a Domain-Specific Language

A domain-specific language to specify AI-specific code smells, combined with static code analysis to detect them automatically in AI codebases.

Diagram: catalog construction, detection approach with SpecDetect4LLM, and assessment of LLM code smells

ICSE 2026, NIER track

Specification and Detection of LLM Code Smells

A catalog of code smells specific to software that integrates large language models, built from the literature, grey literature and empirical data, then detected automatically and assessed for prevalence.

GLiSE pipeline: prompt, LLM-generated queries, web search, extraction, ML classification, grey literature corpus

MSR 2026, Data & Tool Showcase

GLiSE

A Prompt-Driven and ML-Powered Tool for Automated Grey Literature Extraction in Software Engineering

Master Project

BOAM

A Business Oriented identification Approach of Microservices within legacy systems.

ICSOC 2024, LNCS vol. 15405

“BOAM: A Business Oriented Identification Approach of Microservices Within Legacy Systems”, presented at ICSOC 2024 (International Conference on Service-Oriented Computing) and published in Lecture Notes in Computer Science (LNCS, vol. 15405).

In this paper, we introduce BOAM, a hybrid approach that combines source code analysis with business use cases to automatically identify microservices within legacy systems. Our method leverages static and semantic analysis as well as machine learning techniques to deliver relevant and actionable microservice groupings.

Thanks to my co-authors and the research team for this great collaboration!

BOAM approach: use cases preparation, typed services identification, semantic analysis of source code and use cases, clustering into business-oriented microservices Cover of the ICSOC 2024 proceedings, Service-Oriented Computing, LNCS 15405

Papers

Publications

  1. 2026

    Specification and Detection of LLM Code Smells

    Mahmoudi, B., Chenail-Larcher, Z., Moha, N., Stiévenart, Q., & Avellaneda, F.

    ICSE NIERIn Proceedings of the 2026 IEEE/ACM International Conference on Software Engineering, New Ideas and Emerging Results (ICSE NIER 2026)

    DOI arXiv

  2. 2026

    GLiSE: A Prompt-Driven and ML-Powered Tool for Automated Grey Literature Extraction in Software Engineering

    Mahmoudi, B., Chenail-Larcher, Z., Cherief, H. A., Stiévenart, Q., Moha, N., & Avellaneda, F.

    MSRIn Proceedings of the 23rd International Conference on Mining Software Repositories (MSR 2026)

    DOI

  3. 2025

    A Systematic Literature Review of Machine Learning Approaches for Migrating Monolithic Systems to Microservices

    Trabelsi, I., Mahmoudi, B., Minani, J. B., Moha, N., & Guéhéneuc, Y.-G.

    IEEE TSEIEEE Transactions on Software Engineering, 51(11), 2972–2995

    DOI

  4. 2025

    AI-Specific Code Smells: From Specification to Detection

    Mahmoudi, B., Moha, N., Stiévenart, Q., & Avellaneda, F.

    arXivarXiv preprint

    arXiv

  5. 2024

    BOAM: A Business Oriented Identification Approach of Microservices Within Legacy Systems

    Mahmoudi, B., Trabelsi, I., Tamzalit, D., Moha, N., & Guéhéneuc, Y.-G.

    ICSOCIn ICSOC 2024 Workshops, Lecture Notes in Computer Science (Vol. 15405), Springer, pp. 123–137

    DOI

Writing

Blog posts

Articles published on the Ptidej Team Blog.

Presentations

Talks

Slides

JPMS

Java Platform Module System

Research presentations and posters are linked under each project in Research and Master.

Java Platform Module System

JPMS

Contact

Get in touch

Happy to talk about research collaborations, software quality for AI-based systems, or rugby.