How do I actually learn AI/ML deeply enough to build systems (not just follow tutorials)? [D]

via

Machine Learning

I'm stuck in a loop where I consume AI/ML content but can’t move towards actually building real systems. - I understand things at a surface level, but can’t apply them independently - I don’t know how deep I need to go into math, models, or systems - When I try to build something, I get lost between too many topics (ML, LLMs, agents, frameworks, etc.) - I don’t know how to think like a developer/researcher instead of just a user - How do you structure your learning to go from basics → real-world

More from Machine Learning

Visualizing Loss Landscapes of Neural Networks [P]

Machine Learning·12

ICML final decisions rant [D]

Machine Learning·8

Real World Physics-Informed AI Applications [D]

Machine Learning·3

Looking for feedback on OpenVidya: an open-source AI classroom layer for NCERT/CBSE [R]

Machine Learning·8

[D] Self-Promotion Thread

Machine Learning·9

UAI Rebuttal [D]

Machine Learning·9