Nirnay Roy (@nirnayroy1) 's Twitter Profile
Nirnay Roy

@nirnayroy1

Currently into Distributed Systems and Algorithmic Game Theory.
Ex- Data Scientist/AI Engineer/Physicist

ID: 1437635048

calendar_today18-05-2013 07:06:41

229 Tweet

96 Takipçi

1,1K Takip Edilen

Nirnay Roy (@nirnayroy1) 's Twitter Profile Photo

Hi Andy Pavlo (@andypavlo.bsky.social), Do you know why gradient descent does not work very well for query optimization? Is it that computing derivatives are very expensive or more about coming up with objective functions that search the query space exhaustively?

Nirnay Roy (@nirnayroy1) 's Twitter Profile Photo

Tried implementing skip lists in C++ today. With no background in C++, it feels very rough around the edges coming from rust. I am just very used to the borrow checker not letting me shoot myself in the foot. C++ does have ASan, but it's not compile time.

Nirnay Roy (@nirnayroy1) 's Twitter Profile Photo

The intransigent minority rule by Nassim Nicholas Taleb explains the importance of backwards compatibility in software engineering. If a feature is an 80% improvement to the experience but will make 5% devices incompatible, the feature should not be rolled out

Nirnay Roy (@nirnayroy1) 's Twitter Profile Photo

The main reason behind product managers getting paid more than developers is skin in the game. They do not have an upper management to blame, they own features and can be held accountable under all circumstances.

Nirnay Roy (@nirnayroy1) 's Twitter Profile Photo

I have been accepted to School of Solana Season 7 by Ackee Blockchain Security Join me in learning Rust and developing Solana programs! ackee.xyz/school-of-sola…

Nirnay Roy (@nirnayroy1) 's Twitter Profile Photo

A common mistake Data Scientists and Quants make is starting work on a problem without an estimate of the magnitude of returns from investment. A business can potentially pour more resources in operations research and optimization than it gains from that exercise.