SkyGuard.

Developed a YOLO-based computer vision system to detect drones from live camera feeds (static camera or survailence drone feeds). The project focuses on real-time frame processing, server-side handling, and a web interface for visualizing detections, with an emphasis on system design and integration rather than just model training.

Accomplished optimization through reducing the inference cost by increasing the clarity of image sent when low confidence is encountered. (can be further enhanced with auto zoom / autopilot drones)

GitHub (Refer ReadMe.md)

Hyd-DSS.

Developed a comprehensive emergency decision support ecosystem designed to minimize emergency dispatcher information overload and optimize first responder allocation. This dual-sided platform features a citizen-facing mobile-first SOS panel with auto-geolocation and a live map tracking interface. The system leverages the Google Gemini API to triage, classify, and dynamically assess the severity of incoming alerts (police, medical, fire).

The platform demonstrably bridges the critical gap between civilian distress signals and coordinated, data-driven dispatch operations.

Prototype (front-end only)

Chrome Extentions.

Experimented with chrome extentions like website blocking and internal productivity tools.

Includes custom playback features (auto play - pause, picture in picture) & redirects

JalRakshan.

Built an IoT-based emergency vehicular submersion alert system using Arduino, GSM, and water sensors. Designed a detection mechanism that triggers automatic distress alerts when water intrusion is detected.

The system sends GPS coordinates via SMS to predefined contacts, bridging the gap between civilian distress signals and rapid emergency response.

More Coming soon.