VIT Chennai B.Tech Computer Science

Computer science undergraduate building desktop AI tools, web apps, and practical machine learning projects.

I work across TypeScript, Python, Rust, and Java to turn ideas into usable software. My focus is AI-assisted workflows, local-first desktop apps, and machine learning experiments that are clean to run, easy to read, and simple to improve.

01 / EDUCATION

Education

Aug 2024 — May 2028

Vellore Institute of Technology

B.Tech — Computer Science & Engineering

Chennai, India
Current CGPA: 9.26 / 10.00

2024

Srimathi Sundaravalli Memorial School

CBSE Class XII

Alapakkam, Chennai
96.4% (Computer Science: 100/100)

2022

Srimathi Sundaravalli Memorial School

CBSE Class X

Alapakkam, Chennai
96.6% (Mathematics: 100/100)

02 / ENGAGEMENTS

Experience & Leadership

2026

GirlScript Summer of Code

Open Source Contributor

28 merged PRs across 14 open-source repositories — config standards, CI workflows, Docker environments, and contributor tooling.

2024 — Present

Microsoft Innovations Club

Full-Stack Developer

VIT Chennai · Contributing to club web projects, internal tools, APIs, and practical deployment workflows.

2024 — Present

CodeChef VITC Chapter

Technical Volunteer

Supporting algorithmic practice, competitive programming sessions, and campus technical events.

Oct 2025 — Apr 2026

Android Club VITC

Outreach

Reached out to companies and industry contacts to coordinate speakers, sponsorships, and event partnerships for club workshops and hackathons.

03 / EXPERIMENTS

Selected Projects

01

Atheletia

TypeScript · Tauri · NVIDIA NIM

Local-first desktop productivity hub in Tauri + React + Rust with AI chat, notes vault with local retrieval (RAG), activity tracking, calendar, and music — all data stays on-device via Rust IPC + SQLite.

Rust backend handles file I/O, SQLite persistence, and multi-provider AI inference orchestration.

Original Repository · Desktop Workspace GitHub ↗
02

DuckGraph

TypeScript · tree-sitter · SQLite

VS Code extension that parses files with tree-sitter into a SQLite-backed dependency graph, exposed via an MCP server so AI coding agents can query symbols and connections instead of grepping files.

Concurrent file-parsing engine that indexes cross-file symbol relationships and resolves graph queries via an Express MCP server — designed as a distributed graph system with async worker processes.

Original Repository · VS Code Extension GitHub ↗
03

Amorvin Chess AI

Python · PyTorch

AlphaZero-style CNN (12 residual blocks, dual policy/value head) trained on Lichess data via supervised learning, then fine-tuned with self-play RL using MCTS with UCB1 and Dirichlet noise.

Data pipeline converts raw PGN game data into structured board-state tensors for training.

Original Repository · Deep RL + MCTS GitHub ↗
04 / CAPABILITIES

Technical Capabilities

Languages

  • TypeScript
  • Python
  • Rust
  • JavaScript
  • Java
  • SQL
  • C
  • Shell/Bash

Frontend

  • React
  • Next.js
  • Vite
  • Tailwind CSS
  • Framer Motion

Desktop

  • Tauri
  • Rust IPC

Backend

  • Node.js
  • FastAPI
  • Flask
  • MySQL
  • MongoDB

AI / ML

  • PyTorch
  • NVIDIA NIM APIs
  • HuggingFace
  • CUDA Python
05 / CREDENTIALS

Certificates & Course Links

Full Stack Open University of Helsinki
Verify Credential
Verify Credential
Building RAG Agents with LLMs NVIDIA Deep Learning Institute
Verify Credential
Fundamentals of CUDA Python NVIDIA Deep Learning Institute
Verify Credential
Verify Credential
06 / INQUIRIES

Get in touch

I am looking for software engineering internships, open-source mentorship, and project collaborations where I can keep building useful, well-documented tools.