About
I am an early-career researcher working in Artificial Intelligence, Computer Vision, and Autonomous Systems. My background is in Computer Systems Engineering, but my main focus is on formulating clear research questions and designing systems that solve them reliably.
Currently, I am building multi-agent pipelines for automated scientific discovery, exploring ways to evaluate LLM decisions, and using multimodal learning for complex data like physical signal processing. Instead of just treating AI as an application, I want to understand and improve how these systems reason, execute tasks, and verify information in complex environments.
Education
Bachelor of Engineering in Computer Systems Engineering (2022-2026)
Sukkur IBA University
CGPA: 3.80/4.00 | Rank: 2/39
Sukkur IBA University
CGPA: 3.80/4.00 | Rank: 2/39
- Bachelor's Thesis: "Machine Learning-Based Subsurface Defect Detection in CFRP Using Optical Pulsed Thermography"
(Supervisor: Dr. Junaid Ahmed)
Experience & Interests
Independent Researcher
Sukkur IBA University (Aug 2025 - Jul 2026)
Sukkur IBA University (Aug 2025 - Jul 2026)
- Developing execution-grounded multi-agent frameworks to automate research workflows and experimental validation.
- Investigating LLM decision-making and verification techniques to reduce hallucinated outputs in autonomous systems.
- Applying deep learning to multimodal defect detection, signal processing, and computer vision pipelines.
Artificial Intelligence Intern
Mirai School of Technology (Jul 2025 - Aug 2025)
Mirai School of Technology (Jul 2025 - Aug 2025)
- Designed and deployed conversational assistants using agentic AI workflows and large language models.
- Automated data validation and processing pipelines to bridge language models with external APIs.
AI Automation & Multi-Agent Systems
- Automated scientific discovery
- Execution-grounded reasoning
- Multi-agent coordination
LLM Evaluation & Decision Making
- LLM reliability and safety metrics
- Mitigating hallucinations/uncertainty
- Active claim verification
Computational Perception
- Multimodal learning & sensor fusion
- Inverse problems & 3D understanding
- Optimization in dynamic systems
Publications
Computational Depth Measurement in Thermographic Video: Overcoming Spatial Overfitting via Spatio-Temporal Decoupling (2026)
Zain Ul Abidin, Habeeban Memon, Junaid Ahmed.
Under Review in Transactions on Instrumentation and Measurement, IEEE.
arXiv: 2608.29223
Zain Ul Abidin, Habeeban Memon, Junaid Ahmed.
Under Review in Transactions on Instrumentation and Measurement, IEEE.
arXiv: 2608.29223
Three-Channel Thermographic Fusion for Architecture-Independent Defect Detection and Segmentation in Carbon Fibre Reinforced Polymer Composites Using Deep Learning (2026)
Zain Ul Abidin, H. Memon, Junaid Ahmed.
Under Review in Engineering Applications of Artificial Intelligence, Elsevier.
SSRN Preprint
Zain Ul Abidin, H. Memon, Junaid Ahmed.
Under Review in Engineering Applications of Artificial Intelligence, Elsevier.
SSRN Preprint
Debond Detection in CFRP Structures using Optical Pulse Thermography and Deep Learning (2026)
H. Memon, Zain Ul Abidin, Junaid Ahmed, A. Siddiqua.
5th International Conference on Computing, Mathematics and Engineering Technologies (iCoMET), IEEE, pp. 1-6.
DOI: 10.1109/iCoMET69771.2026.11591843
H. Memon, Zain Ul Abidin, Junaid Ahmed, A. Siddiqua.
5th International Conference on Computing, Mathematics and Engineering Technologies (iCoMET), IEEE, pp. 1-6.
DOI: 10.1109/iCoMET69771.2026.11591843
Metal Defect Detection in Pulsed Thermographic Imaging via Temporal Eigen Projection and Robust Low-Rank Tensor Completion Model (2025)
Zain Ul Abidin, Junaid Ahmed, Dong Wang, Guiyun Tian.
Under Review in Nondestructive Testing and Evaluation, Taylor and Francis.
ResearchGate
Zain Ul Abidin, Junaid Ahmed, Dong Wang, Guiyun Tian.
Under Review in Nondestructive Testing and Evaluation, Taylor and Francis.
ResearchGate
Eddy Current Pulsed Thermography based Efficient Defect Detection in Metals via Keyframe Extraction and Accelerated Sparse Decomposition (2025)
Junaid Ahmed, Zain Ul Abidin, Guiyun Tian, Dong Wang.
Under Review in NDT & E International, Elsevier.
ResearchGate
Junaid Ahmed, Zain Ul Abidin, Guiyun Tian, Dong Wang.
Under Review in NDT & E International, Elsevier.
ResearchGate
Mitigation of DDoS Attacks with Machine Learning, Deep Learning, and Transformers (2025)
Zain Ul Abidin, Haseeb, H. Memon, M. Younis, Junaid Ahmed.
International Journal of Multidisciplinary Conference Proceedings (IJMCP), Vol. 2(1).
DOI: 10.61503/Ijmcp.v2i1.181
Zain Ul Abidin, Haseeb, H. Memon, M. Younis, Junaid Ahmed.
International Journal of Multidisciplinary Conference Proceedings (IJMCP), Vol. 2(1).
DOI: 10.61503/Ijmcp.v2i1.181
Research & Engineering Projects
AgentForge: Autonomous Multi-Agent Research Pipeline (2026)
- Engineered a modular framework to autonomously plan, execute, and verify code and experimental research tasks.
- Integrated isolated sandbox testing and an anti-hallucination citation layer to ensure highly reliable outputs.
CMM-DSE: Cross-Modal Consistency-Guided Multimodal Fake News Detection (2026)
- Designed a multimodal architecture leveraging a Cross-Modal Consistency Module to explicitly measure text-image alignment.
- Implemented a Dempster-Shafer evidential fusion layer to output principled uncertainty scores alongside standard predictions.
X-ray Prohibited Item Detection in Airport Inspection (2026)
- Developed an automated computer vision pipeline leveraging deep learning architectures for rapid security screening.
- Segmented prohibited items with high precision within complex, highly cluttered X-ray baggage image scans.
KL-Divergence-Driven Online Sensor Relocation for Barrier Coverage (2026)
- Implemented sliding-window kernel density estimation alongside a greedy sensor relocation algorithm for dynamic tracking.
- Improved the mean detection rate from 65.16% to 87.30% within simulated non-stationary physical environments.
Specializations
- Deep Learning Specialization (Udemy) : CNNs, RNNs, Transformers, GANs, Diffusion Models, and Explainable AI (SHAP, LIME).
View Credential - Machine Learning Specialization (Stanford / Coursera) : Supervised and unsupervised learning, neural networks, and decision trees.
View Credential - Mathematics for ML and Data Science (DeepLearning.AI) : Linear Algebra, Calculus, Probability, and Statistics for ML.
View Credential - Large Language Models (Udemy) : Tokenization, vectorization, fine-tuning, and deploying LLMs.
View Credential - Google Prompting Essentials (Coursera) : Effective prompting, prompt engineering, and LLM workflows.
View Credential
Technical Skills
- Autonomous AI & LLMs: Agentic Workflows (ReAct, Planning) | Prompt Engineering | RAG (ChromaDB, Embeddings) | LLM Evaluation
- Machine Learning: Deep Learning | Ensemble Methods | Computer Vision (Segmentation, Detection) | Tensor Methods
- Programming Languages: Python | C/C++ | MATLAB
- Tools & Environments: PyTorch | Docker | Git | Edge Computing (TensorFlow Lite)