Trustworthy AI and Safety
Designing safety-aware machine learning systems, fail-controlled classifiers, and reliability methods for black-box models.
trustworthy AI, model safety, uncertainty, classifiersBiography
About
AI/ML Engineer and PhD Researcher specializing in machine learning, deep learning, computer vision, and natural language processing. Experienced in LLMs, diffusion models, federated learning, and trustworthy AI systems.
My work focuses on trustworthy AI, safety-aware machine learning, computer vision, federated learning, and scalable AI systems. I design and evaluate models for dependable behavior in safety-critical and real-world environments, with applied experience in PyTorch, TensorFlow, Hugging Face, Django, and data-driven pipelines.
Research
Interests
Computer Vision and Medical Imaging
Developing image analysis systems for detection, classification, person re-identification, and medical imaging applications.
computer vision, medical imaging, detection, classificationLLMs and Generative AI
Building domain-specific LLM workflows, diffusion pipelines, and text-to-image generation systems.
LLMs, diffusion models, generative AI, Hugging FaceFederated Learning
Building privacy-preserving distributed learning systems for image processing and person re-identification.
federated learning, privacy, distributed trainingAI Software Engineering
Developing scalable AI pipelines, model deployment workflows, database-backed systems, and research software.
Python, Django, PyTorch, TensorFlow, data pipelinesEducation
Education
Ph.D. in Computer Science
University of Florence, Florence, Italy
Trustworthy AI, Safety, Machine LearningM.S. in Computer Science
COMSATS University Islamabad, Wah Cantt, Pakistan
Computer Vision, Machine LearningB.S. in Software Engineering
COMSATS University Islamabad, Wah Cantt, Pakistan
Software EngineeringExperience
Academic & Professional Roles
Freelancer - AI & Software Development
Upwork
Data Scientist
AIGOT
Software Developer
Pakistan Ordnance Factories
Research Assistant
COMSATS University Islamabad
Lecturer - Computer Science
Welfare Computer Centre
Technical Profile
Skills
Core Competencies
Programming
Machine Learning
Frameworks
Libraries
Databases
Tools
Credentials
Certifications & Languages
Machine Learning Specialization
Stanford University (Coursera)
Deep Learning Specialization
DeepLearning.AI (Coursera)
Python for Everybody
University of Michigan (Coursera)
Machine Learning with Python
IBM Cognitive Class
Data Science 101
IBM Cognitive Class
Python 101 for Data Science
IBM Cognitive Class
UrduNative
PunjabiNative
EnglishFluent (C2)
ItalianBasic (A2)
Selected Work
Featured Publications
Fail‐Controlled Classifiers: A Swiss‐Army Knife Toward Trustworthy Systems
FA Khokhar, T Zoppi, L Montecchi, A Ceccarelli, A Bondavalli
Projects
Research Software
Food Dish Generation System
A text-to-image generation system for food image synthesis using diffusion models and LLM-generated captions.
Python, TensorFlow, Transformers, LLMs, Hugging FaceFail-Controlled Classifiers
Safety-aware classifiers that can reject uncertain predictions for dependable AI systems.
Python, PyTorch, Scikit-learn, Classification, Trustworthy AIPerson Re-Identification using Federated Learning
A privacy-preserving person re-identification system trained across decentralized datasets.
Python, PyTorch, OpenCV, Federated LearningAwards
Recognition
Best Paper Award
30th IEEE Pacific Rim International Symposium on Dependable Computing (PRDC), Seoul, South Korea
Updates
Latest Notes
Starting My PhD Research Portfolio
A short update about the themes, methods, and direction of my doctoral research.