AARYAN KAKAD

The Lockdown That Changed Everything

Five years ago, the best era of my life started. COVID lockdown. Everyone had terrible experiences - mine was different. I was 12. Playing cards and ludo with my family every day. Then, randomly, I discovered I could learn anything from YouTube. First: investing and stock markets. Then: Python.

I built my first project at 13. The habit of self-learning started in a lockdown. You would have never seen this self-taught guy if that lockdown never existed. Everything happens for a reason - even the bad things. Especially the bad things.

Markets, Code, and the Itch for Business

By 13, I was obsessed with stocks and crypto. I loved predicting what would happen next. At 16, I discovered the online business world - SMMA, dropshipping, all of it. I tried four times. Failed four times. Every single one crashed.

But I learned how businesses actually work from failing that much. I learned what actually matters.

ChatGPT Drops. I Was Just a User.

When ChatGPT dropped, I didn't build it. I was just another user, asking it to fix my code, wondering how the hell this thing worked. For three years, I wondered.

I Stopped Wondering. I Started Building.

In June 2025, I decided to learn ML from scratch. I haven't told anyone in real life what I'm doing. Just Twitter. This has been a solo mission.

Every morning at 5 AM, I wake up and work until 9 or 10. Four to five hours of intense, deep ML work. Then I go to the gym. Every day. I sacrificed my entire social life for this. It's just me, the gym, and ML.

I started with agentic AI and LangChain. Then moved to core ML. I took CS231n - finished every lecture, every assignment. I finally understood how things actually work.

When I get curious about a topic, I watch Karpathy's videos. But here's my real secret: I print research papers. Physically. The ones that make me curious. I read them, understand them, explain them in my own words, and write the explanations directly on the paper itself. That's how I learn now.

My first real build was fine-tuning ResNet18 on 5,500 images to rate faces. When it worked, that was the moment I knew I could actually do this. Then a Siamese network that finds what celebrity you look like.

Then MedVLM - that's the one I'm proud of. A hybrid Vision Transformer that reads chest X-rays. Trained on just 3,400 X-rays, on a free-tier Kaggle GPU. No budget. No lab. Just me and the code.

I'm not doing this to become a machine learning engineer. I'm not trying to get hired at Google to work on someone's recommendation algorithm.

I want to apply ML and build huge companies that solve real problems at scale. I failed at business four times at 16 because I didn't have the leverage. Now I have the leverage. Now it's time to use it.