Curriculum Vitae
Abdul Sami
Machine-Learning Researcher
Summary
Motivated machine learning researcher with a solid foundation in deep learning and generative modeling. Currently specializing in Diffusion Models for image generation, with a strong focus on computer vision and real-world AI applications. Passionate about advancing multimodal generative systems and contributing to impactful, cutting-edge research
🎓 Education
M.Sc. Computer Science – GPA: 4.16/4.5
2023-09 – 2025-08
Soongsil University
B.E. Software Engineering – GPA: 3.73/4.0
2018-11 – 2022-12
Mehran University of Engineering & Technology – SZAB Campus
💼 Work Experience
Research Assistant
2023-09
System Software Lab, Soongsil University
- Conducting advanced research in machine learning and generative modeling, with a focus on diffusion models for image generation and font synthesis.
- Developed a novel few-shot conditional diffusion pipeline for structurally accurate multilingual font generation across Korean, Chinese, and English scripts..
- Trained and evaluated YOLOv8 models for object detection; supported experiments in image segmentation and classification tasks.
- Contributed to multiple publications: one accepted at an international conference (ICOIN 2025), one journal paper under final review, and another in preparation.
- Built and fine-tuned models using PyTorch, incorporating VGG-based feature extraction, CLIP embeddings, OpenCV processing, and reference-style input conditioning.
- Led dataset preparation, model training, and metric-driven evaluation (SSIM, LPIPS, FID) for font generation and vision-related projects.
- Actively collaborated in a fast-paced research environment exploring LLM integration and multimodal generative AI systems.
Teaching Assistant - Artificial Intelligence & Deep Learning
2024-03
Soongsil University
- Supporting graduate-level courses in Artificial Intelligence and Deep Learning for Master’s students.
- Conducting hands-on coding tutorials and lab sessions focused on neural networks, CNNs, and modern deep learning frameworks (e.g., PyTorch).
- Delivering occasional lectures and assisting with conceptual understanding of key topics in generative models and applied AI.
- Designing mid-term and final exams, grading assignments, and providing one-on-one academic guidance to students
🛠 Skills
Programming
- Python
- Java
- C++
- JavaScript
- LaTeX
Deep Learning
- PyTorch
- TensorFlow
- HuggingFace
- Weights & Biases
Generative AI
- Diffusion Models
- GANs
- VAEs
- CLIP
- UNet
- ResNet
- VGG
Computer Vision
- YOLOv8
- Segmentation
- Sobel
- Feature Matching
ML Workflow
- Data Prep
- SSIM
- LPIPS
- FID
- Perceptual Loss
Cloud & DevOps
- GCP Vertex AI
- Docker
- Linux
- Colab
- Kaggle
Tools
- NumPy
- SciPy
- Pandas
- Matplotlib
- SPSS
📚 Publications
Conference paper presenting a text-guided diffusion approach for Korean fonts.
Manuscript on component-level font synthesis.
2025
Proposed a unified diffusion-based framework for multilingual font generation, introducing Multi-Scale Style Infusion for fine-grained style control, Structural Consistency Loss for stroke preservation, and CLIP-based Style Loss to enhance style transfer fidelity.
This thesis focuses on image generation of multilingual characters using diffusion models. By disentangling content and style, the model generates visually consistent and structurally accurate character images across different writing systems..
Presentations
🌐 Languages
English
Fluent
Urdu
Fluent
Sindhi
Native
Korean
Intermediate
🎯 Interests
Human-Centered AI
Cultural Exchange
Language Learning
Traveling
📞 References
Prof. Jaeyoung Choi
Professor & Thesis Advisor, Soongsil University – choi@ssu.ac.kr, +82-10-3311-0684
Prof. Jongsun Choi
Associate Professor, Soongsil University – jongsun.choi@ssu.ac.kr