Hey, I'm Vedant Deshmukh

I build things that think 🤖

I'm a computer engineering student with a habit of turning ideas into experiments. Whether it's crafting a chess AI that thinks 5 moves ahead, training LSTMs to predict weather, or simulating galaxies in C++ — I love digging into how things work and making them work better.

About Me

I'm Vedant — part developer, part curious explorer. I love building intelligent systems, tweaking my dual-boot Arch setup, and pushing limits with code. I enjoy turning abstract ideas into real projects that I can run, test, and iterate on — from AI models to simulation engines.

I've trained a weather prediction model using LSTM and real-world data, simulated the N-body problem with OpenGL, and built a chess engine that learns through self-play (inspired by AlphaZero). I also like hacking together utilities — like a terminal-based anime player, a CLI command helper, and custom F1 tools — just for the joy of it.

My goals? Build systems that learn, reason, and adapt — and hopefully work at DeepMind one day.

Outside code, I'm into F1, story-driven games, and philosophical rabbit holes. I try to stay curious, build honestly, and keep learning.

Projects

WTF! (What's the Function)

A terminal tool that lets you find shell commands using natural language. Built to reduce Google searches and boost flow.

Go

Celestialsim C-GL

A real-time 2D gravitational simulator for the N-body problem. Great for learning physics, math, and graphics the hard way.

C++ & OpenGL

Chess AI (AlphaZero-inspired)

A learning chess engine trained with grandmaster games and self-play. Built with reinforcement learning and MCTS logic.

Python + PyTorch

Weather Predictor & Monitor

Developed an LSTM-based weather prediction model using real-time Meteostat API data — as part of a college PBL project.

Python

Custom F1 Track Editor

Inspired by F1, this private app lets you design and manage your own racing circuits. Think motorsport meets code.

Dart

Highlights & Achievements

⚙️

Weather Prediction Model

Trained an LSTM model with 90%+ accuracy over 7-day forecasts using real-world meteorological data

♟️

Chess AI Engine

Built a learning chess engine using grandmaster PGNs, LSTMs, and Monte Carlo Tree Search

🚀

N-body Physics Simulator

Simulated the N-body problem from scratch with real physics and OpenGL rendering

🐧

Arch Linux Customization

Customized Arch Linux with Hyprland & Sway for full tiling workflow and dev productivity

🏁

F1 Development Tools

Built custom F1 utilities and track editors for fun and learning

Let's Connect

Feel free to reach out!