Hi, my name is Manjot.

I finish a computer science degree at UNLV in December, and I start a master's in machine learning at Georgia Tech in January. Most of what I actually know did not come from coursework. It came from building things end to end and then dealing with whatever broke once other people were using them.

The work I am proudest of is usually not the first version of anything. It is the second one, written after I found out the original approach did not hold up. A recommendation pipeline that recomputed every user's weights and took thirty seconds. A language model that was allowed to decide too much. Those rewrites taught me more than the initial builds did, and they are the parts of these projects I can talk about for an hour.

Outside of that I spent two years as the only teaching assistant for an operating systems course, which mostly meant reading other people's kernel code and figuring out why it deadlocked. I competed with UNLV's ACM chapter and placed first in our Codeathon.