Edge AI
Quantized CNNs on Jetson Orin Nano and Raspberry Pi.
Integrated MCA student and AI & IoT developer building systems for real-world constraints.
I'm Soumya Das. I grew up in a village in West Bengal. My family still lives there.
I'm an Integrated MCA student at Parul University, Vadodara (CGPA 8.54). I also work as an AI & IoT Innovation Intern at Tinkering Hub.
I carry the village with me. Every system I build starts with one question: “Would this work back home?”
I started with small web projects: a marketplace, a college site, and a learning platform. Then I moved to AI and IoT.
I learned one hard thing. A model that needs the cloud fails where the cloud is weak. So I learned to make models small and fast enough to run on the device.
Quantized CNNs on Jetson Orin Nano and Raspberry Pi.
Sensor, model, alert. No cloud needed.
Django, PHP, Node.js, React, MySQL, and MongoDB.
Python scripts and n8n workflows.
Under 100 ms on-device inference.
Built for farmers and used at 3+ farms.
Made for 200+ students.
Includes ML fraud detection.
Edge AI, IoT products, rural or low-connectivity tech, and first-version web apps.
Pure UI styling or native mobile apps. I'll tell you early.
Start with a short note: what problem you have, and who it is for.