I build things across the full spectrum — from web applications to low-level systems, and increasingly into the internals of machine learning.
Currently a B.Tech student at NIT Jalandhar, though most of what I've learned has come from sitting down and building things I didn't fully understand yet. That's still how I prefer to work.
My most substantial project so far is a deep learning framework built from scratch — tensor operations, an autograd engine, neural network modules, and a Python API that mirrors PyTorch, all backed by a C++ core. The goal wasn't to compete with existing libraries but to understand exactly what happens when you call .backward(). That kind of first-principles curiosity is what drives most of what I build.
Beyond that, I've been writing a terminal emulator in Rust, building a health-tech platform, and contributing to open-source projects like Plane and Cal.com — not just using them, but reading the code, understanding the architecture decisions, and contributing back where I can.
I'm drawn to problems that sit at the boundary of systems and product — where understanding the internals actually changes what you can build at the surface. Data structures and algorithms in C++ aren't an academic exercise for me; they're part of how I think about the problems I'm solving.
Open source is something I take seriously. I believe good software should be readable, well-structured, and built in public where possible.