Ayush Sonone

Built a shadow deployment system to validate new application versions without impacting live users. The tool mirrors real production API traffic to a parallel UAT environment and runs both versions in sync. A comparison engine automatically checks HTTP status codes, response payload integrity, and latency deltas to detect functional regressions and performance bottlenecks before rollout.

Implemented with a React-based dashboard for configuration and real-time reporting, and a Java + Spring Boot backend using asynchronous parallel calls for high-throughput testing. Generates structured reports (HTML/CSV/JSON) with mismatch logs, match percentage, and latency analysis, designed to integrate with observability tooling like Prometheus/Grafana for monitoring.

OCR and computer vision pipeline using Python and OpenCV for real-time inventory monitoring with sensors. Reduced warehouse waste by 30% through intelligent stock tracking and automated alerts.

Deployed on DigitalOcean cloud with a scalable architecture. Integrated sensor data with ML models for predictive inventory management and automated restocking recommendations.

Built using React.js and Django with NDVI satellite data visualization and Movebank API integration. Features real-time alerts and animated dashboards for environmental monitoring.

Processes satellite imagery to detect deforestation patterns and tracks wildlife migration using Movebank data. Interactive map visualizations with real-time alerting for environmental anomalies.