The site is new. The thesis is not.

Two years ago, this portfolio was trying very hard to prove that I could build things.

It had a 3D computer, an orbiting planet, animated technology icons, tilting project cards, and a loading screen for the loading screen. It was fun. It was also forty-five megabytes of evidence that I had not yet learned the difference between making something and making something useful.

Then Pulkit and I started building the precursor to Tsenta.

We had applied to more than 3,000 jobs between us. The problem was not abstract. We knew exactly how it felt to lose a weekend to Workday, rewrite the same resume bullet again, and still have no idea whether the application reached a person.

That changed what I wanted a personal site to do.

The work is the proof

I do not need a floating TypeScript logo to say that I write TypeScript. The more interesting evidence is the product, the decisions behind it, and the things that broke along the way.

Tsenta began as automation. The first versions could move through application forms, but they asked users to trust a black box with something deeply personal. That was the wrong product.

We rebuilt it around a simpler idea: people should be able to see, change, pause, and understand what an agent does for them. Transparency was not a settings page. It changed the architecture and the interface.

That version helped take us to Y Combinator.

What belongs here now

This site has three jobs:

Explain what I am building and why.
Keep a selective record of the things that shaped how I build.
Give me a place to think in public.

I plan to write about browser agents, product trust, startup infrastructure, and the economics of software that does real work. I will also write about the less tidy parts of building Tsenta: discarded systems, expensive assumptions, and the decisions that looked obvious only afterward.

The new site is smaller, faster, and easier to change. That is not the interesting part, but it makes the interesting part easier to publish.