Here is another question that I keep getting in various forms:
Every Cloud Platform already has APIs to get everything done. Should I still learn machine learning?
I don’t think I would ever get a chance to build my own models, should I still spend my time learning machine learning?
This question has become very frequent recently and there is a good reason for it. Most of the Cloud service providers have become very aggressive when it comes to providing competing APIs for various AI services such as Speech to Text, Image to Text (OCR), Text to Speech, etc. It gives an impression that this knowledge is very specific to a particular vendor.
The truth is almost every research in Machine Learning is public. Even the most recent breakthroughs such as GPT3 are published and available to all. We are living in a world where researchers publish their work which is globally available. One of the prime reasons for accelerated research is sharing of knowledge. So, no matter which cloud platform you use, you will be using the same internals.
For machine learning, there are various libraries in different languages - all use the same fundamental algorithms. For example, PyTorch and TensorFlow both have the same underlying neural network.
You can use the APIs of Cloud Providers but know the foundation’s concepts would help you do a better job as well as move to different cloud providers when needed.
It is similar to the fact that the philosophy of music is the same no matter which brand of guitar do you use.
I would strongly recommend this curriculum for the holistic learning of Machine Learning and Deep Learning: https://cloudxlab.com/course/12/machine-learning-specialization-includes-deep-learning