Research
Full-stack AI engineer and researcher in computer vision, multimodal deepfake forensics, and PEFT/LoRA. This page collects current research interests, publications and presentations, and ongoing work.
The research agenda connects visual understanding, multimodal forensics, and efficient adaptation of foundation models.
Research outputs currently documented in the portfolio.
Accepted and presented · IBCAST 2026
Dual-stream multimodal deepfake detection with DINOv2, Wav2Vec 2.0, and LoRA. Open-set evaluation on MAVOS-DD; accepted and presented at IBCAST 2026.
View the projectCurrent ongoing work; results and publication details are not yet documented here.
Ongoing research
Ongoing research on open-vocabulary semantic segmentation for transparent objects, using parameter-efficient fine-tuning.
View the projectConcise summaries are drawn directly from the project case studies.
Dual-stream multimodal deepfake detection with DINOv2, Wav2Vec 2.0, and LoRA. Open-set evaluation on MAVOS-DD; accepted and presented at IBCAST 2026.
Ongoing research on open-vocabulary semantic segmentation for transparent objects, using parameter-efficient fine-tuning.
Academic context for the research work, including the thesis supervision documented in the education record.
Sep 2022 – Jun 2026
University of Engineering & Technology, Taxila
Earned a Bachelor's degree in Computer Science with a CGPA of 3.70/4.00 (September 2022–June 2026). Final-year thesis: DIWA-Net — dual-stream multimodal deepfake detection with DINOv2, Wav2Vec 2.0, and LoRA; accepted and presented at IBCAST 2026. Supervised by Dr. Rabbia Mahum (UET Taxila; Researcher at KFUPM). Relevant coursework includes Artificial Neural Networks, Machine Learning, Data Structures and Algorithms, Databases, Operating Systems, Computer Networks, Probability and Statistics, and Linear Algebra. IELTS Academic 7.5.
Education, experience, and technical background in a downloadable format.