cv

Curriculum vitae of Shakeeb Murtaza.

Contact Information

Name Shakeeb Murtaza
Professional Title Postdoctoral Fellow
Email shakeebmurtaza@outlook.com

Professional Summary

Postdoctoral researcher working at the intersection of trustworthy AI (explainability, robustness, evaluation) and applied machine learning, with publications in Pattern Recognition, NeurIPS, and WACV.

Experience

  • 2025 - present

    Montréal, Canada

    Postdoctoral Fellow (Industry Collaboration with Genetec)
    École de Technologie Supérieure (ÉTS) / LIVIA
    Research on explainability and responsible AI for transformer-based models deployed in real-world settings.
    • Lead research on explainability and responsible AI for transformer-based models, from method design to reproducible experiments.
    • Collaborate with academic and industry stakeholders to define research milestones and translate findings into actionable directions.
    • Drive rigorous evaluation with baselines, ablations, and metrics; maintain reproducible experiment pipelines.
    • Co-author manuscripts, contribute to funding and grant materials, and present results at conferences.
    • Mentor students on experimental design, code quality, and reproducible research practices.
    • Manage a high-performance research computing environment (20+ GPU servers) supporting large-scale training.
  • 2021 - 2023

    Montréal, Canada

    Deep Learning Intern (Mitacs)
    Ericsson
    Industrial research internship during the Ph.D. program.
    • Conducted deep-learning research and large-scale distributed training on internal and national GPU clusters (Alliance/Compute Canada).
    • Automated data and experiment pipelines, improving reproducibility, throughput, and stability for computer-vision workloads.
  • 2019 - 2020

    Islamabad, Pakistan

    MS Research Fellow (Medical Imaging, Funded Thesis)
    National Center of Artificial Intelligence (NCAI), COMSATS University Islamabad
    Research in the Medical Imaging and Diagnostic Lab.
    • Conducted medical-imaging research focused on explainability and counterfactual reasoning with deep generative models.
    • Containerized and automated training workflows on research compute clusters for reproducibility and scale.

Education

  • 2020 - 2025

    Montréal, Canada

    Doctorate in Engineering (Ph.D.)
    École de Technologie Supérieure (ÉTS)
    Deep learning with minimal supervision for visual tasks
    • Thesis: Deep Weakly Supervised Learning Networks for Object Localization (https://espace.etsmtl.ca/id/eprint/3975/)
    • Honoured on the ÉTS Honour Roll (Summer 2025) for academic excellence.
  • 2018 - 2020

    Islamabad, Pakistan

    Master of Science
    COMSATS University Islamabad
    Computer Science
    • Specialization in generative deep models for visual tasks.

Awards

Publications

  • 2026
    TeD-Loc: Text Distillation for Weakly Supervised Object Localization
    Under revision at Pattern Recognition
  • 2025
    CoLo-CAM: Class Activation Mapping for Object Co-Localization in Weakly-Labeled Unconstrained Videos
    Pattern Recognition
  • 2024
    SR-CACO-2: A Dataset for Confocal Fluorescence Microscopy Image Super-Resolution
    NeurIPS
  • 2023
    DiPS: Discriminative Pseudo-Label Sampling with Self-Supervised Transformers for Weakly Supervised Object Localization
    Image and Vision Computing

Skills

Trustworthy AI: Model explainability (attribution, concept/prototype methods), robustness, safety-minded evaluation, reproducibility
Machine Learning & Deep Learning: Transformers (ViT/VLM), self- and weakly-supervised learning, domain and test-time adaptation, experimental design, evaluation and metrics
Biomedical Imaging: Histology and microscopy image analysis, weakly supervised localization, dataset publication
Research Computing: Python, PyTorch, Bash, Linux, multi-GPU training, HPC schedulers (Alliance/Compute Canada), Docker, Git, CI/CD

Interests

Research Interests: Weakly supervised learning, explainable and trustworthy AI, vision-language models, person re-identification, biomedical image analysis