Hassan Masroor

Architected end-to-end ML pipelines and built scalable backend inference infrastructure on GCP serving 1,000+ concurrent users, achieving 99.9% uptime and 60% latency reduction via model compression and optimized deployments.

Delivered full-stack AI automation and system design for SaaS platforms, led process digitization for Aviation FOS, and developed edge-optimized CV and NLP solutions that doubled performance in low-resource language tasks through fine-tuning and compression.

Contributed to cyber-physical systems research by developing multi-agent navigation and coordination algorithms ensuring safety compliance and building atmospheric-hydrologic forecasting systems integrating sensor networks with ML models.

Developed neural network compression methods (distillation, pruning, quantization) achieving 70% model size reduction with <5% accuracy loss and built few-shot learning frameworks maintaining 92% accuracy with minimal data.

Activities and societies: Research Assistant at Intelligent Machines Lab; Research & Development Associate at Cyber Physical Systems Lab; projects in neural network compression, few-shot learning, multi-agent systems, and forecasting.

Bachelor of Science in Computer Science completed with coursework and research in ML, computer vision, and systems engineering.