Shehzad S.

Designed and built the core backend architecture that powers ValetEZ’s mobility and parking automation platform, handling 150k+ daily transactions across multiple campuses, tech parks, and real-world environments. Key Contributions - Architected distributed, event-driven backend systems in Node.js & AWS to enable high throughput with minimal latency. - Designed a hybrid sync model using SQS + MQTT + local caching to ensure device reliability even with poor network connectivity. - Implemented microservices for entry/exit validation, parking allocations, Fastag flows, transaction sequencing, audit logs & reconciliations. - Improved system performance and reduced downtime significantly by migrating critical legacy components to modern Node.js services. - Implemented end-to-end observability (error tracking, device health, logs, metrics) to support large-scale deployments. Impact - Scaled from a few thousand to 150k+ daily transactions. - Increased reliability of on-ground IoT systems. - Enabled faster operations, improved throughput and reduced failure rate across all parking sites.

Built a robust ANPR system combining on-edge inference, 4K image capture, vision models, and cloud-based verification to automate vehicle entry/exit and parking workflows. Key Contributions - Designed real-time image capture pipeline from 4K cameras with optimized cropping & scaling logic. - Integrated YOLO-based vehicle + plate detection with a custom post-processing stack for accuracy improvements. - Built a hybrid inference model using local inference + cloud verification to maximize accuracy in diverse lighting conditions. - Developed high-performance Python service on Odroid C4 to orchestrate detection, preprocessing, quality filtering, queueing, and cloud sync. - Built PHP APIs to receive and validate ANPR events in real-time with the parking backend through SQS. Impact - Achieved significantly improved plate recognition accuracy. - Reduced manual validation load for parking teams. - Enabled automation-ready entry/exit systems and improved throughput during peak hours.