Pranav Deshpande
What I do
I’m a Research Lead at JPMorganChase, where I build AI systems for finance - the kind where models have to actually work or you lose money. I’ve invented 10+ U.S. patents and published research in leading conferences and journals.
I studied computer science at NYU and engineering at NIT Nagpur, India. Before moving into research, I spent years in industry learning the hard way what works when you put models in front of real users.
What I keep coming back to
AI is changing how science works. The old loop - observe, hypothesize, experiment, publish - assumed human cognition was the only engine of discovery. That assumption is dissolving. AI can now generate hypotheses, synthesize entire literatures, and design experiments, which means humans are becoming directors of the search rather than the ones doing all the searching.
Small teams are getting absurdly powerful. AI multiplies what each person can do, so tiny groups can now build what once required large organizations. That changes startups, governments, and everything in between.
The bottleneck is shifting from execution to judgment. When AI handles the doing, the hard part becomes deciding which problems matter, what counts as a good outcome, and when to stop iterating. That’s harder to automate than it looks.
What I’m writing
Build Your Own AI is a book I wrote for people who weren’t satisfied just using AI and want to understand how it actually works. I was one of them. I started as a software developer who could build things, but wanted to understand how AI actually worked - beneath the APIs, beneath the frameworks, down to the actual ideas and math. After years of self-study, experiments, and a lot of trial and error, I crossed over from writing software to doing AI research. The book is my attempt to shorten that path for others: no API calls, no fine-tuning someone else’s model - you build the thing layer by layer and understand why it works while you do it.
AI Systems and the Future of Civilization is a series of essays about how AI restructures science, organizations, finance, governance, and humanity itself.
About this site
This is where I think out loud about AI - what it actually is, how it works, and what it’s doing to the world. The essays are long-form pieces on the structural stuff: how scientific discovery is changing, why small teams are becoming absurdly powerful, what it means for humans when execution gets cheap and judgment becomes the scarce resource.
If you’re trying to understand where this technology is actually going - not the press releases, not the doomscrolling - this is the place.