At Gujarat Technological University-School of engineering and Technology, I developed a privacy-preserving threat detection pipeline for distributed IoT sensor networks. It reduced per-round communication overhead by 97.3% compared to FedAvg.
I used Python, TensorFlow, and TensorFlow Federated to run decentralized machine learning, then containerized the simulation with Docker and Docker Compose so edge nodes could operate without exposing raw sensor data.
For an ESP32 posture detection project, I paired an MPU6050 motion sensor with MQTT telemetry and connected the hardware to ThingsBoard for dashboards, alerts, and real-time notifications.
My work also includes MATLAB simulation of F-OFDM waveforms, where I demonstrated up to 46% throughput gain over conventional 4G LTE OFDM architectures. As a Frontend Developer Intern at DS Software, I co-developed an educational web application using HTML, CSS, and JavaScript.
