Developed a gesture recognition model using TinyML to detect hand motions (swipe, wave, tap) from accelerometer and gyroscope data on an Arduino Nano 33 BLE Sense.
Trained and optimized a TensorFlow Lite model for on-device inference, enabling real-time, low-power classification without cloud dependency.
Applied the system to EdTech use cases such as contactless classroom interaction (e.g., navigating slides, triggering learning modules) and assistive technology for students with limited mobility.
Optimized model with quantization and pruning, achieving >90% accuracy while running on constrained microcontrollers.
Integrated results with MQTT-based dashboard to log usage, track learning engagement, and demonstrate potential for smart classrooms.
Project Head
04.2025 - Current
Developed an IoT-based dust monitoring system using Arduino sensors to measure particulate matter (PM2.5/PM10), temperature, and humidity in real-time.
Improved environmental monitoring on construction sites by providing real-time insights, enabling proactive health & safety measures.
Integrated a Telegram Bot to send real-time alerts and daily summaries directly to contractors, ensuring immediate action when dust levels exceeded safety thresholds.
Education
B.Tech - Industrial Internet of Things
Vivekananda Institute of Professional Studies - Technical Campus, GGSIPU