About Me

AI & Full-Stack Engineering

I'm Shuvankar Debnath — a Software & AI Engineer based in Kolkata, India, building where intelligent systems meet modern web. With hands-on experience across the MERN stack and applied AI, I design and ship production-grade applications that solve real problems, not just proof-of-concept demos. During my internship at IEMRF, I contributed directly to live backend systems — building REST APIs and MongoDB schemas for ERP and LMS platforms used by real users. My engineering philosophy is simple: systems should be scalable, maintainable, and built with intent.

Computer Vision & Applied AI

My deepest technical work lives at the intersection of deep learning and real-world data. I trained a multi-generation plant disease detection pipeline — benchmarking YOLOv8, YOLOv9, and YOLOv11 on a custom Roboflow-annotated dataset — achieving approximately 92% mAP. This wasn't academic; every version was trained, evaluated, and compared to understand what actually changed across model generations.

Building AI Products

Note Loom is my flagship project — a full-stack AI-powered educational SaaS platform built on the MERN stack with role-based access control (RBAC) and LLM integration. It's not a tutorial clone; it's a system architected for real multi-user use, with distinct access flows for students and admins, an AI layer for personalised learning experiences, and a production-ready backend. Building it taught me how to think about AI features as infrastructure — not just API calls bolted onto a UI.

Electrical Engineering Roots

I'm pursuing a B.Tech in Electrical Engineering at IEM Kolkata (graduating 2027), and that background shapes how I think about software. Control systems, signal processing, and systems modelling give me a different mental model for designing software — one that treats architecture like a feedback loop: inputs, state, outputs, and correction mechanisms. My EV Battery MPC Simulation project, which implements Model Predictive Control for battery management, is a direct expression of that: EE theory translated into a software simulation environment using Python and MATLAB.

Video Editing & Graphic Design

Outside of engineering, I take on freelance work in video editing and graphic design — two crafts I've developed seriously over the years. On the video side, I work across DaVinci Resolve, Adobe Premiere Pro, Wondershare Filmora, and CapCut, handling everything from cinematic edits and YouTube content to client deliverables. I'm comfortable with colour grading, motion pacing, and audio sync — the details that separate polished work from rushed cuts. For design, I use Adobe Photoshop, Adobe Illustrator, and Canva to create brand assets, social media graphics, and marketing materials. Both disciplines feed directly into how I think about UI and product presentation in my engineering work.

Content & Community

Beyond code, I create technical content under the identity SUVNKR. On YouTube (@ShuvankarTechie) I share project walkthroughs, AI/ML deep-dives, and engineering breakdowns aimed at students transitioning into AI roles. I believe in learning in public — building things, documenting them, and putting the work out there for others who are on the same path.