Projects

Born from a love for galaxies and the reality of wrestling with a Dobsonian telescope in the dark. Manually calculating coordinates and fumbling with star charts gets old fast when you're trying to find M31 at 2 AM. So I built a mobile companion that does the math for you — real-time GPS positioning, compass guidance, and Alt-Az alignment. Now the night sky feels magical again, not like a trigonometry exam.

When someone's passion for cricket meets someone else's love for software, you get this. What started as conversations about weekend leagues turned into building an entire platform for managing teams, tracking stats, and coordinating matches. Rode the wave of the 2024 T20 World Cup all the way to the US market. Turns out the best collaborations happen when you find your co-founder on the pitch.

My new favorite way to spend free time — which inevitably led me back to my computer desk. Practice, practice, practice takes hours, so why not track it? Built to log range sessions, measure progress, and turn swings into data. Turns out even when I'm trying to relax, I end up building software. Some habits are hard to break.

Started with shameless inspiration from Wordle, then struggled to add my own creative spin. The goal? An English learning tool for myself and anyone else trying to expand their vocabulary without feeling like they're studying. Six attempts. One hidden word. A daily habit that feels less like work and more like play. Still iterating, still learning.

Built this as an employee at Octo.AI, then circumstances led to open-sourcing it. Not the path we planned, but honestly? Got more work, learned more lessons, and the project reached more people. Sometimes the best outcome isn't the one you expected. A big data analytics framework designed to process streams without drowning in complexity.

A college side project born from eavesdropping on PhD-level materials science lectures and borrowing knowledge from a professor generous enough to humor an undergrad. Computer vision meets metallurgy — teaching machines to identify material composition pixel by pixel. Built with OpenCV when I barely understood either domain. Sometimes the best learning happens when you're in over your head.