Islamabad, Islāmābād, Pakistan
I lead AI initiatives focused on building production AI systems that solve real-world problems. My work sits at the intersection of Generative AI, AI Agents, and Computer Vision, where the goal is not just to experiment with models, but to deploy scalable systems that organizations can rely on. As Head of AI at Red Buffer, I work with startups, enterprises, and government organizations to design and deploy intelligent systems that automate workflows, extract insights from complex data, and power new digital products. This includes building AI copilots, document intelligence platforms, computer vision-based monitoring systems, and large-scale machine learning pipelines. Over the past 9+ years, I’ve worked on end-to-end AI platforms, from research and model development to production deployment. My experience includes multi-GPU training pipelines, scalable MLOps infrastructure, real-time vision systems, and AI-driven automation platforms running in cloud and edge environments. Alongside industry work, I am also a researcher in biomedical imaging, working on AI methods for giga-pixel histopathology analysis and cancer diagnosis. My research focuses on deep learning approaches for nuclei detection, segmentation, and cellular community analysis in digital pathology. I’m particularly interested in building AI systems that move beyond prototypes into reliable production environments, helping organizations unlock meaningful value from artificial intelligence. You can also explore the recommendations on my profile, where colleagues and collaborators have shared their experiences working with me. Note: The views expressed on this platform are my own and may not necessarily reflect my company's or its officials' views.
-Developed Patient Survival Analysis pipeline using Whole Slide Images and Genomic data. -Developed NC-Net model that performs Nuclei Instance Segmentation for Whole Slide Images. -Developed Semantic Segmentation Pipeline for MoNuSeg Histopathology Dataset -Developed Semantic Segmentation Pipeline for CPM15, CPM17, and CoNSeP Dataset -Developed Instance Segmenetaion Pipeline for PanNuke Dataset
- Orchestrated Multi-Note, Multi-GPU training pipeline for training 5 Billion Image Dataset using AWS SageMaker - Built and Orchestrated a highly scalable Cloud Cost Anomaly Detection pipeline for Yotascale - Built Computer Vision based Safety App for an Oil Drilling Ship using Video Streams that raise real-time alerts - Developed POC for Detection of Money Laundering using Transactions Data - Developed Property Detection System for Automation of Property Tax Collection - Developed Baseball Card Matching and Meta-Data Extraction Pipeline - Developed Mobile Activity Classification Pipeline for Multiple Apps using Video Streams
- Orchestrated Multi-Note, Multi-GPU training pipeline for training 5 Billion Image Dataset using AWS SageMaker