The End

In the very beginning of this guide, I presented you with some performance gobbledigook, which you may or may not have understood. What about now?

Image credit

Thank you for sticking with it and getting all the way to the end of this. Take a moment to enjoy the fact that you accomplished something -

🌌 Queue Eric Idle! 🌌

In case you feel like learning more; where to from here?

Before I answer that, just one aside. It is very easy to just absolutely wear yourself out doing courses like these online in your free time. Look out for yourself. You won't learn much if you're burnt out. I have been there more than once. Please take care of yourself!

Machine Learning Operations

If you feel like you dipped your toes into efficiency and you want to keep dipping your toes into new stuff that will make you an overall better engineer, I can't recommend Nicki Skafte Detlefsen's course Machine Learning Operations enough. In fact, I ripped off his website to create this one (with his permission). In it you will learn software engineering principles applicable to AI development such as repoducibility and deployment. It's also my recommendation which contains more memes than this site. Which is something we should all aspire to.

Machine Learning Systems

If you want to learn more about optimizing machine learning systems - this online book written, edited and curated by Vijay Janapa Reddi looks to be a very extensive resource, still under rapid development.

Visual Computing Systems

Kayvon Fatahalian has a wonderful course up on visual computing systems which combines topics from graphics, computer vision, image analysis and deep learning. Specifically, focused on the systems used in those fields. You can the most recent version here. The topics included seem to change from year to year. It can definitely be worth it to go back and see which topics were included in earlier years, through his teaching website. It's worth noting however, that this course is fairly advanced and if you want to do the exercises you should probably be familiar with C++.

Parallel Computing

In the same vein, Kayvon Fatahalian and Kunle Olukotun have a course on parallel computing. If you want a more thorough and academically rigorous introduction to parallelism, GPU's and memory hierarchies it is available here. Do note that you probably need to be familiar with C++ to do the exercises.

TinyML and Efficient Deep Learning Computing

If you are more into video resources, the course TinyML and Efficient Deep Learning Computing by Song Han contains both exercises, slides and videos for a wide look at more efficient deep learning.