AI/ML is not one monolith, and there are many different levels of knowing it:
-
AI user - learn how to do prompt engineering with LLMs and tools like DALLE & SORA
-
AI integration developer - call REST APIs of ML models to use AI within your application
-
Junior ML engineer - Knows how to train ML models using Python
-
Senior ML engineer - Knows how to deploy and run ML models in production on Cloud, and build end-to-end ML training and deployment pipelines
-
ML researcher - instead of just using existing ML algorithms, can do research to improve them or invent new algorithms, and publish papers for them
Also, there are multiple sub-levels within each level, so just under level 3 (junior ML engineer), you can have these various skills to learn:
3a. Knowledge of classic ML algorithms for classification and regression with libraries like scikit-learn
3b. Knowledge of Neural Networks & Deep Learning with libraries like TensorFlow & PyTorch
3c. Knowledge of how to apply ML & DL to different areas like image processing, audio processing, video processing, time series
3d. Understanding of fundamentals of Natural Language Processing (NLP)
3e. Knowledge of how Transformers & LLMs work, and how to do fine-tuning & RAG with LLMs etc.
Any of the people in the various levels above can be said to have "learned AI/ML".
So the real answer is, there is no fixed timeline. You start from scratch and you can keep learning more and more about AI/ML, and keep becoming more knowledgeable about it your whole life, as you get deeper and deeper into it, like any other field in the world.
So you decide, what level of knowledge you're targeting, and depending upon your target level, it can take from months, to years, to decades to reach. Good luck ππ½