About Me

Hey, I’m Om — a Machine Learning enthusiast from New Delhi, India.

I’m currently an undergrad at GGSIPU majoring in IT, and most of my time goes into building, breaking, and rebuilding machine learning systems. I’m the kind of person who learns by doing… then doing it again the right way.

So far, I’ve been experimenting and building across Computer Vision and Agentic AI. Projects like VisionSynth and InboxPilot started as ideas, turned into messy prototypes, and slowly became working systems. ML is the core. Everything else is just supporting machinery. Nothing fancy. Just me trying to understand how intelligence actually works under the hood.

I’ve previously worked with the Indian Navy (WESEE) as an ML intern. I spent two months working on real-world, defense-grade ML problems. Easily one of the coolest experiences I’ve had so far.

What I’m doing right now

• Going deeper into core ML. Working with advanced models and improving classical algorithms through feature engineering, tuning, and careful evaluation

• Tinkering with Agentic AI systems. Building them, breaking them, and figuring out where autonomy actually helps and where it’s just noise

• Sharpening my math foundations that sit under ML and decision-making. Less “exam math”, more “why does this actually work” math

• Staying consistent on Kaggle through competitions and shared notebooks. Mostly to build intuition. Grandmaster title would be a nice side effect

• Taking a small detour into psychology. Part curiosity, part pattern recognition, part trying to understand the humans behind the data

How I got here

I didn’t start with a grand plan. I’m learning by building systems, watching them fail, and correcting for it. Progress, when it happens, comes from repetition, not insight. One project led to another. One idea led to another. Somewhere along the way, I realized I enjoy making systems that think.

I’m not the smartest person in the room. I’m just comfortable being wrong repeatedly until something works.

Outside of tech

When I’m not building or debugging, I read, cycle, and learn for no immediate payoff. It helps me slow down, think clearly, and avoid mistaking constant output for progress.

If you’re building something interesting

I’m always open to working with people who care deeply about what they’re building and are willing to iterate until it works. If that sounds like you, let’s talk.