CV
Download the PDF for the full CV.
Contact Information
| Name | Ali Hussain |
| Professional Title | Undergraduate Researcher · PhD applicant (Fall 2027) |
| ahussain.bscs23seecs@seecs.edu.pk |
Professional Summary
Undergraduate researcher working on visual representation learning, multimodal learning, efficient foundation models, robust computer vision and document intelligence. First-author ICDAR 2026 paper and equal-contribution BMVC 2026 paper.
Experience
-
2024 - Present Islamabad, Pakistan
Undergraduate Researcher, Handwriting Recognition
TUKL-NUST R&D Center, NUST
With Prof. Faisal Shafait. Led the research end to end, through the manuscript, for a first-author ICDAR 2026 paper.
- Built the OUHD-L online Urdu handwriting dataset (2,403 lines, 311 writers) with coordinates, pressure, tilt and timing.
- Designed multimodal fusion of online trajectories with offline visual representations; reduced CER from 4.21% to 3.66%.
-
2026 - 2026 Kaiserslautern, Germany
Research Intern, Efficient Vision & Edge AI
RPTU Kaiserslautern-Landau
Final-year-project research with Dr. Christian Weis on foundation-model distillation and edge deployment.
- Implemented YOLO baselines and evaluated frozen teacher encoders with linear heads for pest detection and classification.
- Ran inference on Raspberry Pi hardware and measured power and deployment performance.
-
2025 - 2025 Kaiserslautern, Germany
Research Intern, Document Image Forensics
German Research Center for Artificial Intelligence (DFKI)
With Prof. Andreas Dengel and Dr. Sheraz Ahmed.
- Led benchmark design, dataset sourcing and experiment implementation for studying JPEG-induced structural bias in document-tampering detection.
- Designed controlled evaluations across detectors, datasets, JPEG settings and manipulation alignments (PyTorch, OpenCV).
Education
-
2023 - 2027 Islamabad, Pakistan
B.S. (expected 2027)
National University of Sciences and Technology (NUST)
Computer Science
- Relevant coursework: Algorithms, Artificial Intelligence, Probability & Statistics, Machine Learning, Deep Learning.
-
2021 - 2023 Multan, Pakistan
Beaconhouse College Program
GCE A-Levels
- Gold Medalist; 100% scholarship holder.
Publications
-
2026 Mistaking Periodicity for Manipulation: JPEG-Induced Structural Bias in Document Tampering Detection
British Machine Vision Conference (BMVC)
Two query-free attacks, Grid-Aligned Forgery and Pad–Recompress–Crop, reveal that state-of-the-art document-tampering detectors depend on JPEG 8×8 grid statistics.
-
2026 Online Urdu Text-Line Recognition by Bridging Stroke Dynamics and Offline Representations
International Conference on Document Analysis and Recognition (ICDAR)
New OUHD-L dataset (2,403 lines, 311 writers) and online-feature fusion into pretrained offline recognizers; CER 4.21% → 3.66%.
Awards
-
2023 High School Gold Medalist and 100% Scholarship
Beaconhouse College Program
Skills
Programming: Python, C/C++, JavaScript
Machine Learning: PyTorch, Hugging Face Transformers, scikit-learn, knowledge distillation, LoRA/QLoRA
Vision & Multimodal: OpenCV, detection, tracking, OCR/HTR, vision-language models, diffusion models
Systems: Linux, Git, Slurm/HPC, Docker, Raspberry Pi inference and power measurement