Machine Learning from First Principles | Massachusetts Institute of Technology Research Experience | Math, Theory & Practice

Learning Python internals with Kumar completely changed how I think about Python. He doesn’t just explain what happens under the hood; he helps you understand why it happens, step by step, until concepts that once seemed intimidating start to make sense.

What stood out to me most was his ability to break down complex ideas without sacrificing depth. Instead of asking you to memorize how things work, he helps you build the mental models to reason about them yourself.

If you’re someone who wants to move beyond writing Python code and truly understand the language, I’d highly recommend learning from Kumar. His approach makes the journey into Python internals both approachable and genuinely rewarding.

I’ve been learning Machine Learning from Kumar for the past six months, and it has been one of the most valuable learning experiences I’ve had. His biggest strength is that he doesn’t just teach you how something works; he makes you understand why it works, the mathematics behind it, and what is happening under the hood.

Over these six months, we’ve explored the mathematical foundations of ML, linear algebra, calculus, probability, statistics, optimisation, loss functions, gradient descent, backpropagation, neural networks, and the principles behind deep learning. What makes his teaching exceptional is how he connects mathematical derivations with algorithmic intuition, implementation, and real-world applications. You don’t just learn to use libraries or reproduce tutorials; you develop the ability to reason about models independently.

The most tangible outcome of this journey was landing an internship at Google. The conceptual clarity and problem-solving mindset I developed through his teaching helped me approach challenging technical questions with much greater confidence. Rather than preparing me merely to pass interviews, Kumar helped me build an understanding that I could actually apply to unfamiliar problems.

His background in ML research at MIT, competitive programming, and software engineering is reflected in the rigour of his teaching. He can move from mathematical proofs to implementation details without losing sight of the bigger picture. He challenges you to think deeply, welcomes difficult questions, and patiently approaches a concept from different angles until it clicks. He also adapts his teaching to your level, making even intimidating topics approachable without oversimplifying them.

Whether you’re a complete beginner, a student strengthening your fundamentals, a software engineer transitioning into ML, or someone aspiring to work in AI research, Kumar can help you build a much deeper understanding of the field. If you want to go beyond blindly applying PyTorch or TensorFlow and genuinely understand how and why machine learning algorithms work, I cannot recommend him highly enough.

If you’re looking for a mentor who combines mathematical depth, practical engineering, first-principles thinking, and a genuine commitment to your progress, Kumar is the person to learn from.