AI/ML Engineering student | Builder | Community Leader

B.Tech CSE (AI/ML) @ UPES Dehradun · Class of 2028

Currently interning at NewCycl — building Rawbin, an electric home composting device — while pursuing the Data Science & AI Program by IIT Roorkee E&ICT Academy (March–November 2026).

I like building things that combine technical depth with real-world usability: ML systems, full-stack platforms, and the communities that grow around them.

• MASS — Multi-Agent Startup Simulator — building a multi-agent system where specialized AI agents (CEO, Marketing, Finance, Product, Sales) collaborate to turn a single startup idea into a full business strategy
• System Design + DSA in Java — learning scalable architecture patterns and sharpening algorithmic thinking through hands-on Java implementation
• Full-Stack Development — building real projects with and without AI assistance, covering both MERN and modern Next.js patterns

Languages

Frameworks & Libraries

AI / ML

Databases & Cloud

Tools

A multi-agent system that simulates a real startup team from a single idea. Specialized agents — CEO, Marketing, Finance, Product, and Sales — debate and coordinate to produce a complete business strategy including plans, pricing, roadmaps, and go-to-market output. Built as a hands-on experiment in multi-agent architecture and systems thinking.

Python Multi-Agent Systems GenAI

Freelance project — production-ready ordering platform for a live food brand. Features real-time order tracking, push notifications, a dual-menu system with automatic stall activation during college events, and a full admin dashboard. Has processed 100+ live orders without major operational issues.

React TypeScript Supabase Capacitor Tailwind Vite

Clean, minimal portfolio built with Next.js 14 (App Router), Tailwind CSS, and Shadcn UI. Features a serverless markdown blog via gray-matter and next-mdx-remote, a contact form via Resend, and full responsive + dark mode support. Deployed on Vercel.

Next.js TypeScript Tailwind Shadcn UI Resend

Physics-aware Discrete Event Simulation framework using SimPy that models Slotted ALOHA and CSMA/CA protocols. A Random Forest classifier dynamically selects the optimal MAC protocol in real time. Reduced collision rates to approximately 0.29% under 17,000+ packets/sec stress tests.

Python SimPy Random Forest scikit-learn

End-to-end ML pipeline classifying Delhi NCR Air Quality Index into health-risk categories using XGBoost. Improved accuracy from ~57% to 98% through data harmonization and hyperparameter optimization. Includes an interactive CLI for real-time prediction.

Python XGBoost scikit-learn Pandas

Algorithmic trading bot for the IMC Prosperity challenge. Implements safe mid-price handling, dynamic execution thresholds, and queue-priority logic for profitable trade execution.

Python Algorithmic Trading

Custom 2D game engine built from scratch featuring a complete Entity Component System, batch rendering with OpenGL, an integrated Dear ImGui editor with docking, JBox2D physics, and GSON-based scene serialization.

Java OpenGL JBox2D

• Intern @ NewCycl — building Rawbin, an electric home composting IoT device (June 2026)
• PR & Sponsorship Head — UPES IET Community + IEEE Computational Intelligence Society Student Chapter
• Lead Enabler & Senior Mentor — managing ~100 students across SAMARTH and SRIJAN programs at UPES
• IIT Roorkee E&ICT Academy — Data Science & AI Program (March–November 2026)
• Head Boy — Model School Rohtak, representing 1,160+ students (2023–2024)