Mitish Raina

Thinking about building something new this weekend.. it's been a while since I built something crazy' good.. any problem y'all facing?

To be in my presence is a blessing.

@NoahKingJr Call boy

Software Engineers what's your plan B if AI replaces you?

What's more important in clg? Attendance or skills? Cause my fuckass clg wants to pay student with good grades because "they deserve it".. (iss professor or uske 75+ wale students.. dono ki MAA KI CHUT)

@UsmanReads Yes and I want someone who can give cs.lg endorsement on arxiv

Dear Algorithm Can you please connect me with someone who can give me arxiv cs.AI endorsement?

Me when exams(side thing) start to interrupt my daily 9-5 routine(coding)😔

@DevByLakshya Is it more like AI AI internship (MLE) or just calling some APIs?

Are you interested in a remote AI internship? 📌 PPO (Pre-Placement Offer) based on performance

Can't wait for WW3, WW3.o and WW3.5 mini

@AnujS21202 Bhai gift krdo mujhe mac 🙄🤷

NEVER BUY A MACBOOK IN 2026 NEVER BUY A MACBOOK IN 2026 NEVER BUY A MACBOOK IN 2026 NEVER BUY A MACBOOK IN 2026 NEVER BUY A MACBOOK IN 2026 NEVER BUY A MACBOOK IN 2026 NEVER BUY A MACBOOK IN 2026 #MacBookAir #MacBookPro #AppleLaunch #Apple #AppleMusic

@LowKeyDevs Bhai mere ghar bhi order krde ek

@codewithdev @codegirl007 Is there any way I can get one mac?🤷

@codegirl007 I have 3 MacBooks, 2 Ipads but never got a chamce to giveaway. I only use 1 macbook rest of the devices are just in cooling period 😭

I'm wiping my 5 year old Mac Book Pro clean and giving it away. I don't need 2 laptops. Runs great, no damage. Primarily used for dev work in Neovim when traveling. If interested, send me a DM. I will cover shipping if it's "within reason" (tbd what "within reason" means).

@mehuljindal18 Can I apply? I'm a prefinal year student🙄

🚀 We’re building Aurora AI (stealth) at the intersection of AI × offline world Ex-YC founders (built → scaled → sold) Backed by Global VCs & Angels Hiring Founding Engineers who want: ⚡️ high ownership • 🧠 deep code • 😈 zero BS Know some1? Tag them & win Airpods Max🎧

Implemented almost the whole DEEP LEARNING from MLP, CNN, RNN, LSTM to GRU ... Every single concept with only maths.. no libraries.. This one was actually crazy. Abstracted every single thing just to understand what's under the hood. Checkout repo here: gitHub.com/mitishraina/Ba…

@lossfunk Hey lossfunk, I have doubt, saw your page regarding research intern that you guys hire. One thing that i wanna ask is if I apply, I need to come up with my own idea for research or you guys have a team working around any topic and I can contribute to it?

🥳🎉 Very happy to announce that our papers got into four ICLR 2026 workshops! • VerifAI workshop • I Can’t Believe It’s Not Better workshop • World Models workshop • Test Time Updates workshop See you guys in Rio de Janeiro!

@shookscooks I mean you miss every shot you don't take lmao.. gg man atleast you got the chance. If not a problem can you tell who he is

chief ai scientist at a very big lab/company in india. Just do things anon. Just do it.

From last 2-3 hours, I've been implementing every sinlge optimizer from scratch, every single activation function from scratch.. and the clarity i have now is insane like i can imagine its working even while sleeping lol Checkout repo: github.com/mitishraina/Ba… And use define by run, and you'll actually process how algorithms are designed.

@MankyDankyBanky @_reesechong x.com/i/status/20281… Checkout this for ground up building

I built a neural network from scratch without using PyTorch, TensorFlow, or any libraries for that matter. Instead, I implemented the core math myself. I'm working on making my own machine learning framework from scratch with my friend @_reesechong. He previously trained a similar neural network, but using just scalars. The next step was to use tensors instead. The benefit of this is clear: when using scalars each data point is looped through separately creating its own node in a computation graph. With tensors, these separate nodes are stored together, allowing one forward and backward pass for the whole batch, greatly improving the efficiency of training. We often hear the saying "don't reinvent the wheel", but in my experience rebuilding technologies that abstract away a lot of complexity gives you a better and more thorough understanding of how the system works. Results of the training are shown below. Feel free to checkout the repo and read through the code, linked in the replies.

Implementing algorithms using core maths only, abstracts away a lot of complexities and helps you to visualise working under the hood. Without relying on libraries and building from the ground up is going to take you far more than just working with libraries only. Great work man. One tip is keep a mental model using define by run and DRY principle. And then implement it.