CV

The complete CV can be downloaded from the top pdf button.

Basics

Name Muhammad Osama Zeeshan
Label Computer Vision and Machine Learning Researcher
Email osamaz.oz31@gmail.com
Phone +1-438-357-6482
Url https://osamazeeshan.github.io/
Summary Computer Vision and Machine Learning researcher with a Ph.D. in Systems Engineering at École de technologie supérieure (ÉTS). My research focuses on human-centered visual intelligence, multimodal learning, vision-language models, facial dynamics, and personalized model adaptation under real-world distribution shifts. I also bring more than five years of industry engineering experience in software development, image analysis, and video-processing systems. My research has been published in ECCV, ICLR, WACV, and CVPR Workshops, and I am particularly interested in digital humans, neural representations, and AI-native visual content creation.

Work

  • 2026.02 - 2026.04
    Visiting Research Scientist
    STARS, Inria - French National Institute for Research in Digital Science and Technology
    Valbonne, France
    • Conducted collaborative research on computationally efficient machine learning methods for personalized human behavior and expression analysis.
    • Investigated test-time multimodal domain adaptation methods that improve robustness and efficiency without requiring access to original training data.
    • Explored personalized adaptation strategies for models operating under changing human behavior, subtle facial expressions, and real-world distribution shifts.
  • 2022.01 - Present
    Machine Learning Researcher
    LIVIA & ILLS Labs - École de technologie supérieure (ÉTS)
    Montreal, Canada
    • Developed personalized deep learning models for behavior and emotion recognition, adapting to individual-specific patterns and variability.
    • Designed and implemented unsupervised, gradual, multi-source, and test-time domain adaptation methods for robust generalization under distribution shifts.
    • Built multimodal learning approaches integrating visual, physiological, and audio signals for adaptive prediction and behavior analysis.
    • Conducted large-scale experimental evaluations on noisy real-world datasets to assess robustness, cross-domain performance, and reproducibility.
    • Created Python and PyTorch pipelines for preprocessing, model training, evaluation, ablation studies, and robustness analysis.
    • Collaborated with interdisciplinary researchers on personalized and multimodal AI systems across behavioral health, computer vision, and affective computing.
    • Co-developed the BAH dataset for ambivalence and hesitancy recognition, enabling research in personalization, zero-shot learning, and domain adaptation.
  • 2019.01 - 2022.01
    Senior Software Engineer
    Jin Technologies
    • Developed scalable backend systems and RESTful APIs supporting data-driven production applications.
    • Collaborated with cross-functional teams to design and implement reliable software systems.
    • Optimized systems for performance, reliability, and scalability across deployment environments.
  • 2017.09 - 2019.01
    Software Developer - Video Content Retrieval
    Bahria University Research Lab
    • Developed computer vision pipelines for detecting and recognizing text in video streams.
    • Built systems for indexing and retrieving visual information from large multimedia datasets.
    • Designed data-processing workflows for noisy, low-quality, real-world visual data.
  • 2015.02 - 2017.08
    Android Developer & Image Analyst
    Steprobotics
    • Developed computer vision algorithms for solar irradiance, skyline, and shading estimation.
    • Built the Step Solar mobile application integrating geospatial analysis, local data storage, and external APIs.
  • 2014.01 - 2015.03
    Machine Learning Research Assistant - Computational Biology
    Air University
    Islamabad, Pakistan
    • Developed machine learning and image-analysis pipelines for microscopy-based sperm cell classification and recognition.
    • Processed high-resolution sperm microscopy images and extracted discriminative visual features to identify head and tail abnormalities according to WHO criteria.
    • Developed and evaluated classification models for computer-aided semen analysis across Normal, Abnormal, and Neutral categories.

Education

Publications

Awards

  • 2025
    FRQNT Award
    Fonds de recherche du Québec - Nature et technologies
    Personalized Deep Learning Models for Diverse Individuals, Quebec, Canada.
  • 2022
    4-Year Ph.D. Scholarship
    École de technologie supérieure (ÉTS)
    Fully funded research position at ÉTS, Montreal.
  • 2019
    Top 100 - Hello Tomorrow Global Summit
    Hello Tomorrow
    Recognized at Le Centquatre, Paris, France.

Skills

Programming
Python
SQL
Java
Machine Learning & AI
Computational Efficiency
Model Complexity Analysis
Computer Vision
Multimodal Learning
Self-Supervised Learning
Domain Adaptation
Test-Time Adaptation
Generative Models
Evaluation & Benchmarking
Frameworks & Tools
PyTorch
TensorFlow
OpenCV
Git
Linux
Research Engineering
Reproducible Pipelines
Heterogeneous Data Preprocessing
Large-Scale Experiments
Ablation Studies
Benchmarking