Hey, I'm Aditya.
I build products — and before that, I try hard to understand the person I'm building for. Six years across edtech, impact tech, and consumer; currently product at ConveGenius.AI, mostly pointed at AI.
What's stayed constant isn't a domain — it's a question: how do people decide, learn, and grow, and what does a product owe them? I chase it with equal parts building and research, and I'm suspicious of features that arrive before the problem is understood.
These days that's AI-native product work, a newsletter on the tooling, short reels on the human side of it, and mentoring early-career PMs. If any of that's relevant, say hi.
The arc, drawn out.
Five roles, four kinds of work, one throughline. Move across the years — the curve fills in behind the point and the story changes as you go, including the messy in-between of each switch.
Gujarat’s first Vidya Samiksha Kendra, scholarship bots for 1.5M students across 60,000 schools, a CET bot handling 12 lakh registrations. Public-sector product at population scale.
The common thread is products that meet people where they are — not where a roadmap wishes they were. The current chapter is AI-native product work: what changes when models become a first-class part of the stack, and where I can help PMs and founders navigate that shift without losing the plot.
When I'm not at a screen I'm probably out walking, or in a coffee shop in a city I haven't been to before.
I start with the problem, not the feature.
The building is the visible half. The other half is research — sitting with a problem long enough to model it before writing a line of product. My most recent deep dive: India's career-alignment crisis.
Information ≠ outcome. 90% of Indian students choose a path blind — not for lack of Google, but for lack of an occupational identity to process advice through.
A search engine gives data; people need a mirror. So I framed the missing “operating system” as three measurable pillars — the same ones career-construction research maps to a set of readiness skills.
The way I work a problem is fairly consistent:
- 01Ground it in theory. Career Construction Theory, RIASEC + Big Five, the NACE framework — so the design rests on something sturdier than a hunch.
- 02Validate with users. Interviews until the personas fall out of the data — here, two axes: how well someone can articulate, and how well they know themselves.
- 03Build for the real context. India-native, not a translated Western template: a 1:10,000 counsellor gap, Tier-2/3 realities, image-first over text-heavy tests.
That research became a product — an AI career counsellor ↗ that guides Socratically instead of spoon-feeding, so students build the confidence, not just collect the answer.
Where it's landed.
I think out loud.
Writing and the reels are the reflection loop — where the building and the research get metabolised into something shareable. The rest of me lives here.
Work
Products, side projects & client builds.
Tech Therapy
Short reels where AI concepts explain being human.
Newsletter
Weekly-ish notes from the notebook.
GitHub
Repos, experiments & works in progress.
Geo-Traveller
The travel journal — places & routes.
Let's talk
Mentoring, consulting, or just to say hi.