About Abdul Rehman
AI/ML engineer and computer vision specialist with published medical imaging research, production deployment experience, and active HEC-funded research in large language model systems.
Education
MS Data Science & Machine Learning
FAST National University of Computer and Emerging Sciences (FAST-NUCES)
Graduate program focused on machine learning theory, deep learning, computer vision, and statistical methods. Thesis work on medical image classification with transformer architectures and model calibration.
Research & Publications
Frontiers in Neuroscience (2026)
Calibration-aware CNN-Swin model for cross-dataset brain MRI classification
Published research on brain tumor MRI classification achieving 99.55% accuracy using hybrid DeiT + SVM pipeline. Work addresses model calibration, cross-dataset validation, and uncertainty quantification critical for clinical AI systems.
ICoDT2 (December 2025)
LLM-assisted academic assignment generation and evaluation
Accepted poster on fine-tuned language models for learning-outcome-aligned assessment generation. Research explores LoRA/QLoRA fine-tuning, semantic evaluation, LLM-as-judge methods, and human-in-the-loop workflows.
HEC-Funded Research (Ongoing)
LoRA and multi-agent LLM systems
Higher Education Commission of Pakistan funded research investigating parameter-efficient fine-tuning methods and multi-agent coordination architectures for specialized language model tasks.
Federated Learning Research
Privacy-preserving learning for medical image classification
Research comparing FedAvg, FedProx, FedBN, SCAFFOLD, and MOON algorithms for cervical cell classification across distributed non-IID datasets. Includes client-level explainability analysis for clinical transparency.
Production Experience
Face Liveness Detection System
Developed and deployed biometric verification application for government pension program. System verified more than 40,000 pensioners through mobile-based face liveness detection, handling real-world conditions and diverse user populations.
Mobile & Full-Stack Engineering
Cross-platform mobile development (React Native), backend API design, cloud deployment (AWS), and operational tooling. Experience spans consumer applications (100,000+ users), public sector systems, and operational platforms.
Applied AI Integration
Integration of computer vision, document processing, and ML workflows into production systems. Focus on measurable value delivery, evaluation frameworks, and systems that perform reliably in operational environments.
Technical Expertise
Computer Vision
- Medical image classification
- Face liveness detection
- Vision transformers (Swin, DeiT)
- CNNs and hybrid architectures
- Model calibration
- Explainability (XAI)
Deep Learning
- PyTorch, TensorFlow
- Federated learning
- Multi-seed evaluation
- Hyperparameter optimization
- Model compression
- Mobile deployment
LLMs & Generative AI
- LoRA/QLoRA fine-tuning
- Llama, Mistral models
- RAG systems
- Prompt engineering
- Multi-agent architectures
- LLM evaluation
Full-Stack Development
- React Native
- Backend APIs
- AWS deployment
- Database design
- Mobile-to-operations workflows
- Production support
Connect
Approach
Research informs production work. Methods proven in academic settings—multi-seed evaluation, calibration analysis, ablation studies, reproducible experimental design—transfer directly to commercial AI systems requiring reliability and transparency.
Senior involvement means the person discussing architecture remains involved in implementation. Technical decisions, delivery risks, and product trade-offs are documented and visible rather than hidden in layers of account management.
Discuss Your AI Challenge
If you have an AI/ML problem in medical imaging, computer vision, generative AI, or need production deployment of intelligent systems, we can assess technical feasibility and propose a grounded approach.
Start a Conversation →