At GreenAI, I lead the Data & AI team, working across model design, data pipelines, and repository standards from development to deployment. I built multimodal crop classification models that identify major Pakistani Kharif and Rabi crops with 87%+ accuracy.
I also built a yield estimation pipeline combining Sentinel-1 SAR and Sentinel-2 optical data to forecast district- and farm-level yields before harvest. For flood monitoring, I developed a weekly Sentinel-1 SAR microservice and a labeling and auto-retraining dashboard for field and AI teams.
My work also includes turning satellite data into natural-language diagnostics with spatial feature extraction and RAG-based LLM workflows. Previously, as a Deep Learning Research Intern at Machine Vision and Intelligent Systems Lab, I developed semi-supervised contrastive learning methods and published research in IEEE Xplore.
