How should software engineers learn AI?

Before jumping into all these,please learn linear algebra very well(matrix manipulation, SVDs).your everyday data are not just row/column vectors), basic calculus (you’ll need the chain rule to understand backprop).Build simple networks b4 jumping into transformers & all these

The order matters a lot. Engineers who skip straight to Agents without understanding how attention works end up cargo-culting prompts and wondering why context degrades past 8k tokens. Transformers first, then RAG. Agents without that foundation are just prompt gymnastics.

All these topics are present as really beautiful blogs at

get hands-on with projects; theory's not enough without real-world experience.

because the hard part was never just learning the algorithms, it's applying them to a real problem and making sure the solution doesn't break in production

Learn calculus and linear algebra first

Bhai Apka AI engineering course se chutiya kat gya🥲 Usme kuchh tha hi nhi

Jst three things only???

@st4ayh4rd Do it and then let me know :)

Ok sir, will update u after completion 👍

It’s only 15%

Crazy that ik almost 90% of the above

i would start by rebuilding small pieces, not taking another survey course. one toy search, one eval, one messy agent loop. the gaps show up much faster when something has to run.