kaushik raj

Been building Applicable for last 2month , A platform for IOT enthusiastic! We are pushing code every day , we are trying out everything which can help us to gain the trust of the user ! That it ! #Indianstartup @elonmusk @nikhilkamathcio @Hiteshdotcom @StartupGrind

@Unnati_twts girl writing code 🥸🥸 🚩

Most devs learn theory. Very few build real systems. Launching a 7-day bootcamp to learn System Design + Backend + AI integration by building a real project You’ll go from idea → scalable system → deployed product Starts: 17 June 2026 DM for more info!

@AtharvaXDevs howwwww

I am begging you to take twitter premium and start posting tech and genuine tweets 😭🙏🏻

"A loonshot is something nobody has dared to do before. A goal so ambitious that you don’t even know if its possible." -- naman pushp of airbound (substack, 2022) i am 23 now. i have one skill, it's the ability to tell stories. i want to talk about indians building on research.

@Hikari_07_jp HBD

I turned 19 today. Lately, I've realized there's so much I don't know and so much I need to learn. I'll continue researching local LLMs and sharing useful information🔥 Open source will always win🚀🚀

@sbhm_prksh @zomato congratulations

Spent all my primary and secondary school years here. It was in this same room where I dreamed of becoming a Nobel Laureate. Life had other plans. The journey led me to IIT Mandi. Now, I will be joining @zomato this month as an SDE. Excited for the next chapter. Byebye!

@usman_akinyemi mention some of them 🤌🏻🤌🏻🤌🏻

The aura of Super30 seniors is so unmatched. Those guys are on another level bro.

@Prnvshrma1 which uni bruhhh ?

Bhenchod meri university me apple store khulgya🤣

@Punisher278710 trueee

Every bug teaches a lesson. Some lessons just take 6 hours to learn :(

@0xeik @zeeg same qs why ?

@zeeg why lol ? why spin up a server when you can just run a lambda function for a small service ?

serverless was a mistake

@izzHanu you guys have opening for intern ? Would love to work with you guys 🌸✨

@ShriKaranHanda @OpenAI what it include?

Another delivery from @OpenAI I think

@hitesh_ml recipe?

Made lemon paneer today. Don't go by the looks but it was so damn tasty😋 and 40 grams of protein 💪💪

@NatyaCodes We are in the same boat theapplicable.in check this out and lets connect we guys can build more nd more ✨

hello people of X, i am natya, born and raised in delhi. been doing electronics since i was 10 and programming since i was 14. right now i'm building AI inference on small-scale FPGAs and microcontrollers for edge AI in robots. here's how i got here

@bit2swaz @pixperk dont apply to indian company first rule of software engineering 🌸🤌🏻

@pixperk had enough of this bs :/

mkc agar hire nhi karna toh assignment kyu karwaate ho

@arvind_iiserm Congratulations sir ✨✨✨🌸 People like you make this country great ✨

Congratulations to Gayatri (my joint PhD student with Kavita) for the successful completion of the PhD viva. She has moved to a tech job in a quantum startup in Bangalore.

@aarushidotexe 🍻

got my first PR merged yesterday saw my code live on the website today feels so good 🫶

@kunalvg Indont think so ! They are too closed source and need to open first !

Sarvam will be India’s DeepSeek. I am not an oracle. But everything they have shipped, and the way they have shipped it, makes me bullish. They will thrive.

@Bhakti8997925 build drone virtually first and ship it

Why are drone parts soo costlyyy....whats a broke clg student supposed to do

@TensorTonic i habe trying lot getting stuck due to your vibe coder ui and it frustrating and there lot more bug in the page itself

@ironrobot10 @prathamgrv Auhu waiting for further discussion on this topic ✨

comparing it with leetcode, scaling stateless cpu execution is pretty much a solved problem with serverless and docker, but doing it for gpu sounds incredibly tough. i know about modal and a few other solutions out there, but what did you actually use to make it work? did you build custom containers or use an existing platform to keep it economically feasible?

the amount of systems knowledge I gained from this is genuinely unmatched. the sheer learnings from scaling real GPU workloads for 40k users can't be taught in any course. we gave every user a safe sandboxed container with deterministic execution against test cases on actual GPU hardware. there is no LLM-as-a-judge in the loop that's gonna verify the user's code. the whole thing scales because there is zero shared state between submissions. per language container images means a CUDA kernel submission can't see a Triton Submissions's state. the autoscaling on demand with per second billing instead of reserved GPUs is the only reason this is economically viable solo.