Hey, I'm Parth Batra.
I teach machines to think for a living, and argue about music for free.
I grew up in Gurgaon - close enough to Delhi to inherit its opinions, far enough to pretend I don't live there — and I've spent about four years now somewhere in the messy middle of data, ai, and the question of what any of it actually means.
These days I'm a Senior ML Engineer at Procol, turning supplychain chaos into systems that quietly make decisions. Before that I picked up scars - the kind you only earn by shipping something that breaks at 2am - at Maruti-Suzuki Genpact MunchStudio
But I'm rarely still for long! Sometimes work keeps me in the air - but the trips I actually count are the ones I choose: one big escape a month, the higher the altitude or the colder the better.
I write here to mostly think out loud and part hot takes I'd defend over a drink (like "humour is just how you bleed without anyone seeing the wound"). It's unedited on purpose: a journal that accidently became public.
This place has a few rooms...
blog is the long-form stuff, the things I actually sit down to write; thegoodstuff is everything I'd vouch for - handpicked; quickwits catches the bite-sized raw, rapid-fire, straight off my X feed; workbench is where half-built side projects and learning logs live; list100 is my century of dreams I'm foolish enough to track in public;
Off the clock, you will find me acquiring some new hobby or reading about some niche rabbithole and hunt for artists before spotify's algorithm catches on.
The credentials, for the curious: a dual degree from BITS Pilani—B.E. (Hons.) in Mech and M.S. (Hons.) in Mathematics, which is a fancy way of saying I like building things and proving they work. Less officially, I've logged 100k+ minutes on Spotify 3 years in a row, a statistic I'm arguably prouder of.
For the record, on the day job: I call myself a full-stack ML engineer because I genuinely think i am Mr. Do It All
- AI agents that might just flirt with turing test
- SQL elaborate enough to deserve their own repos
- models that earn their keep instead of decorating a slide
- microservices that i deploy on a friday evening without flinching
- data pipelines that survive contact with real data
Here's to the ones who dream
Foolish as they may seem
Here's to the hearts that break
Here's to the mess we make
If any of that sounds like your kind of noise, you know where the writing is.