Noor Ullah

I am a Machine Learning Engineer and Software Developer based in Islamabad, Pakistan, working across computer vision, deep learning, and full stack systems engineering.

Research Experience

Most recently, I directed machine learning research for a British Council funded collaboration between Quaid-i-Azam University and two UK institutions, the University of West London and the University of East London. I designed the study methodology, coordinated data collection across research teams, and led the machine learning analysis underpinning the project's core deliverables.

Industry Experience

At Kosmic AI, I architected and deployed the backend infrastructure powering an AI video generation platform, integrating third party APIs to support hundreds of concurrent generation requests. I designed the marketplace matching algorithm that improved resource allocation efficiency, and I rebuilt the CI/CD pipeline, cutting release cycle time by 27 percent through database query tuning, backend profiling, and automated testing gates. I partnered directly with Product and Design leadership to translate AI capabilities into shipped features.

Technical Foundation

My technical foundation spans object detection, image segmentation, and image classification, built on supervised and deep learning methods using PyTorch and TensorFlow, with OpenCV, NumPy, Pandas, and Scikit-learn supporting data processing and experimentation.

Education

I am completing an MPhil in Computer Science at Quaid-i-Azam University, following a BSCS in Computer Science from PMAS Arid Agriculture University with a CGPA of 3.67.

Certifications

• AI For Everyone, DeepLearning.AI

• Digital Information Literacy, State University of New York (SUNY)

• Introduction to Mobile Development, Meta

• Technical Support Fundamentals, Google

Beyond Engineering

Outside of engineering, I build custom PCs, compete in chess, and hike. Different pursuits, same discipline: patience with hard problems until they resolve.