Shubhanshu Dubey

WORK HISTORY

EDUCATION

ABOUT SHUBHANSHU DUBEY

I am a software developer and AI enthusiast driven by the challenge of building high-concurrency systems and localized AI solutions. Currently pursuing a dual degree—a B.Tech in AI & ML at Newton School of Technology and a BS in Data Science at IIT Madras—I bridge the gap between rigorous academic theory and production-ready engineering. My technical journey is defined by three core pillars:Generative AI & LLMs: I specialize in orchestrating multi-agent systems and optimizing Local LLMs. Most recently, I developed Nexus-AI, a multi-agent hub using FastAPI and CrewAI that leverages Groq LPUs for sub-second inference. I also led a national-level team for SIH 2025, where we engineered an offline RAG-based AI tutor capable of running on low-resource devices (4GB RAM) for rural education. Full-Stack Engineering: I build immersive, high-performance web applications using the MERN stack, Next.js, and TypeScript. My focus is on creating seamless user experiences with advanced animations (GSAP) and secure, scalable backends. Problem Solving: With an AIR 63 in the ISI entrance exam and a 900+ rating on Codeforces, I have a deep-rooted passion for Data Structures and Algorithms. Beyond coding, I’ve honed my leadership skills serving as a Regional Coordinator for IIT Madras, managing operations for the Mumbai region. I am always eager to collaborate on open-source projects, GenAI innovations, or full-stack challenges. Technical Toolkit:Languages: JavaScript, TypeScript, Python, SQL. Frameworks/Libraries: React, Next.js, Node.js, Express, FastAPI, Tailwind CSS. AI/ML: CrewAI, RAG, Hugging Face, OpenAI API, Qwen/Llama models. Tools: Prisma, MySQL, MongoDB, Git, Docker, Postman.