Every government keeps two sets of books:what it promises, and what it builds.I study the distance between the two —the exact point where policy breaks.
Parth Gupta — Berlin, raised in Silicon Valley. Researcher, summit organizer, podcast host, and builder of Canopy. This site is the paper trail.
If you only have three minutes
Parth Gupta is a Berlin-based IB Diploma student (Class of 2027, born and raised in Silicon Valley) working at the intersection of AI governance, trade policy, and sustainable development. His research includes an Extended Essay on the regulatory limits of facial-recognition markets, an SSRN preprint testing whether US export controls accelerated Chinese supply-chain independence, and a logistic-regression model of G20 trade retaliation. He built and runs Canopy, an AI-powered platform generating daily policy briefings across 193 UN member states, and has placed in national and regional essay competitions. Outside research, he organizes policy summits — including a 100-person climate summit — serves on the US Embassy Berlin-Brandenburg Youth Advisory Council, hosts two podcasts, and edits his school newspaper.
00 · Preface
A short letter before the exhibits.
Berlin, 2026
Dear reader,
I grew up in Silicon Valley and moved to Berlin in 2021. I've lived in the space between those two places ever since — the Bay Area taught me what technology believes about its own power; Europe taught me what happens when institutions try to govern that power from the outside. Almost everything I'm curious about lives in that gap.
The official version: I write research on AI and trade policy, organize summits, host two podcasts, edit my school paper, and built a platform that checks whether 193 governments actually keep their promises.
The unofficial version: ten years at a piano, a ukulele I'm still bad at, a badminton habit I refuse to quit, and a school of seventy-plus nationalities, which taught me that most interesting things start as a translation problem.
I learn things by building them, and I change my mind a lot — usually because someone in the room knew something I didn't.
What follows is evidence, not advertisement.
— Parth
The exhibits.
Six bodies of work. One question: why doesn't policy do what it says it will?
Research
I read the fine print so I can argue with it.
An Extended Essay on why facial-recognition markets defeat regulation · an SSRN preprint on how US sanctions made China more self-sufficient · a logistic-regression model of trade retaliation.
Practice
Then I go watch it break in person.
A G20 think-tank summit from the staff side · a 100-person climate summit I organized and co-moderated · the US Embassy's youth council in Berlin · Model UN · a Duke of Edinburgh expedition.
Media
Then I talk to people about all of it.
Two podcasts, 28 episodes — one about young people doing uncommon things, one about the economics underneath the news · Editor-in-Chief of The Paw Print · essays on Medium.
Canopyupdates itself daily
And when the tool I needed didn't exist, I built it.
AI-written briefings on 193 countries' development promises, generated fresh every morning. The only exhibit that changed since you opened this page.
Essays
Then I compress it into a 2,000-word argument.
Six competition essays — four John Locke Institute submissions, a regional-winning HCGEC argument on mandatory voting, and a nationally-placed USA Economics Olympiad essay on the price of hyper-specialization.
Now
What's actually happening this month.
Current work, upcoming commitments, and the things that keep me grounded when the deadlines stack up.
Open questions.
Things my own work raised and didn't settle. If you can answer one of these, the contact form isn't decorative.
Q·01
Can you regulate a market when the buyer can't see the product?
Left over from the Extended Essay — the information asymmetry never resolved.
Q·02
If export controls make their target self-sufficient, what exactly did they control?
Left over from the preprint — 90% localization says the policy worked in reverse.
Q·03
Why do climate pledges survive contact with everything except budgets?
Left over from the climate summit — four experts, no consensus, one good map of the dispute.
Q·04
Does watching 193 governments every day change what they do — or just record what they don't?
The question Canopy can't answer about itself.
If any of this sounds like a conversation you'd want to have, I'd love to have it.
Say hello