What are the next steps I should take after finishing ML on coursera?

Geron's book is good but the first half of it should be stuff that you already know from Coursera. Maybe try the PyTorch book - https://pytorch.org/assets/deep-learning/Deep-Learning-with-...

"Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow" (2019) 2nd ed. by Geron is rated 5-stars with 1723 reviews on Amazon, with good reason.

I think Andrew Ng's course on Coursera is targeted towards a different audience. You are pretty good with math and I think you'll appreciate his CS229 Stanford course more, as he goes through the math, and does right off the bat on the blackboard/whiteboard. He goes covers gradient descent in the first lecture [after the intro lecture]. 2008[0] and 2018[1] versions.

Yaser Abu-Mostafa has CS156 at Caltech, a course and a textbook called "Learning from Data"[2][3].

It depends where you want to go. If you want to learn Python libraries like Tensorflow and Pytorch, Keras, and the like... there are many books for that.

"Deep Learning with Python", by François Chollet.[4][5]

"Data Science from Scratch: First Principles with Python", by Joel Grus.[6][7]

There are the fast.ai courses on https://fast.ai and also have a book titled:

"Deep Learning for Coders with Fastai and PyTorch: AI Applications Without a PhD" by Jeremy Howard and Sylvain Gugger. They have the lectures for several years on machine learning and deep learning.

There are many books and resources on how to train models. However, there are extremely few that enable you to do ML in a non toy project setting and do useful things in the real world.

One recent course is CS329s "Machine Learning Systems"[8] that fits nicely after CS229. I only skimmed some slides, but it appears to be interesting.

Disclaimer: Yes many publications about deployment, and ML project lifecycle management and platforms, but I have blocked most of them, including the largest publication about data science and machine learning on Medium from search results in my browser, because it publishes garbage posts from people who seem to never have worked on a project with actual stakes.

I know it was not your question, but I guess you'd like to do something with ML in the real world. There is a range of skills and process necessary to make things happen. Identifying the right problem, designing the right metrics (tricky, especially when working with a project involving "intangibles"). Reframing. These also require interviewing skills and deep dives into a subject matter (so the ability to discover and curate the right resources, then read and synthesize) to be able to interact with domain experts and see a project through.

- [0]: https://www.youtube.com/playlist?list=PLA89DCFA6ADACE599

- [1]: https://www.youtube.com/playlist?list=PLoROMvodv4rMiGQp3WXSh...

- [2]: https://www.youtube.com/playlist?list=PLD63A284B7615313A

- [3]: https://work.caltech.edu/telecourse

- [4]: https://www.manning.com/books/deep-learning-with-python

- [5]: https://github.com/fchollet/deep-learning-with-python-notebo...

- [6]: https://www.oreilly.com/library/view/data-science-from/97814...

- [7]: https://github.com/joelgrus/data-science-from-scratch

- [8]: https://stanford-cs329s.github.io/syllabus.html