Welch Labs

Welch Labs is made by Stephen Welch, who holds engineering degrees from Georgia Tech and UC Berkeley. He started the channel in 2014 to explain neural networks more clearly than the resources he could find at the time. He has also written an illustrated guide to AI.

The videos explain maths, science and machine learning with detailed animations and a strong sense of story. Topics range from imaginary numbers and the maths of the Manhattan Project to how image recognition and large language models work. Many videos are part of multi-episode series that build an idea step by step.

Some highlights include:

The story of AlexNet, the image recognition system that marked a turning point in modern AI, and what it reveals about how neural networks learn.

The calculations made during the Manhattan Project to check whether an atomic bomb could set the atmosphere alight.

The first part of a series showing why imaginary numbers are a real and useful part of maths.

The channel has over 900,000 subscribers.

AI and machine learning now appear in almost every branch of engineering, from robotics to medical imaging. This channel explains the maths behind them without skipping the hard parts. It suits students heading towards computer science, electronic engineering or a maths-heavy engineering degree, and shows how ideas from complex numbers to calculus are used in real systems.