Me

[The Sigma Mindset @thesigmamindset]

This video will change your life ‼️

[Sann @san_x_m]

His name is Soumith Chintala.

He grew up in Hyderabad and went to a good school, but he was bad at mathematics. He studied at VIT in Vellore, a college that does not carry the prestige of an IIT.

When he finished his degree and applied to twelve universities in the United States for a master’s programme, every single one rejected him, despite a strong GRE score.

He went to the United States anyway.

He applied again, this time to fifteen universities. All rejected him except two.

In 2010, he joined New York University. There he met a professor named Yann LeCun, who was not yet famous, and began working in his lab on a then unfashionable field called deep learning.

The rejections did not stop.

He applied to DeepMind and was rejected. He applied for jobs across the industry and was rejected.

The only offer he received was for a test engineering role at Amazon.

At one point, a visa technicality nearly forced him to leave the United States. He spent months obtaining a waiver so he could continue his work.

He kept building things in the open and sharing his code freely.

In 2016, while working at Meta, he and a small team released an open-source framework called PyTorch.

It was a tool for building artificial intelligence systems.

Today, PyTorch is one of the foundations of modern AI. A significant share of AI research and products around the world is built using the framework he helped create.

He spent eleven years at Meta before leaving in November 2025 to join a new AI lab.

He was rejected by twelve universities, then fourteen more, then DeepMind, then almost every employer he approached.

The tool he helped build is now used by much of the field that once said no to him.

Follow for stories India deserves to remember.

[Srishti @srishticodes]

Want to become an AI Developer in 2026 but confused about what to learn and in what order?

Here’s a clear, no-fluff roadmap people actually follow 👇
1. Foundations (don’t skip this)
• Python (write real code, not just notebooks)
• Data Structures & Algorithms (arrays → trees → graphs)
• Math that actually matters:
– Linear Algebra (vectors, matrices, eigen intuition)
– Probability & Statistics (distributions, expectation, variance)
– Calculus (gradients, optimization basics)

2. Core Machine Learning
• Supervised vs Unsupervised learning
• Regression, Classification, Clustering
• Bias–variance tradeoff
• Feature engineering & model evaluation
• Libraries: NumPy, Pandas, scikit-learn

3. Deep Learning (modern AI starts here)
• Neural Networks & backprop
• CNNs for vision
• RNNs → Attention → Transformers
• Training tricks, overfitting, scaling
• Frameworks: PyTorch (preferred), TensorFlow

4. Specialize (choose ONE, not all) • NLP
• Computer Vision
• Recommendation Systems
• Speech
• Generative AI (LLMs, Diffusion, Agents)

5. Production & Industry Skills
• Data pipelines & APIs
• Model serving & monitoring
• MLOps, retraining, data drift
• Basic system design for ML systems
AI careers aren’t built by chasing trends.
They’re built by strong fundamentals + one deep specialization.

Bookmark this. Repost it.
Your future self will thank you.

[Me @nil24A]

@youtubemusic

[Me @nil24A]

@Google @YouTube @YouTubeIndia @GoogleIndia

YouTube Music needs cross-device queue sync ASAP 🙂
Building the perfect vibe on my phone and then switching to PC to a completely different mood is just painful.