A.6 Frequently Asked Questions

Use this page when the question is emotional or vague. Convert it into a next action.

Fast answers

Question	Short answer	Next action
I am weak at math. Can I learn AI?	Yes. Learn math through code and projects.	Run Python, NumPy, and small ML examples first.
Do I need a GPU?	Not at the beginning.	Use CPU/API first; rent cloud GPU only when training needs it.
Do I need every chapter?	Keep the main path, then choose direction modules.	Finish foundations, then pick RAG/Agent/CV/NLP/multimodal focus.
How many hours per week?	Consistency beats bursts.	4-10 hours/week is enough if sustained.
When should I do projects?	As early as possible, but keep them small.	Build one input -> one process -> one output.
I cannot read papers. Is that fatal?	No. Papers are later supplements.	Read tutorial, code, then paper.
When can I job hunt?	When you can explain 2-3 projects clearly.	Prepare README, metrics, failure cases, and interview story.

Minimum learning path for application builders

Tools→Python→Data→Deep Learning Basics→LLM Principles→RAG→Agent

If you want stronger model foundations:

Tools→Python→Data→Math→ML→Deep Learning→LLM→RAG / Agent

Weekly rhythm

Weekly time	Good rhythm
4-6 hours	2 study sessions + 1 coding session
7-10 hours	3 study sessions + 2 coding sessions
12-18 hours	Add one project/review block, avoid burnout

Confidence reset

When you feel slow, ask:

1. Can I understand more code than last month?
2. Can I modify examples instead of only copying?
3. Can I explain one concept more clearly?

If yes, you are progressing.

Three abilities to strengthen first

1. Python and debugging.
2. Data processing and visualization.
3. Minimal closed-loop project thinking.

Many beginners think they need a more advanced model. Often they first need clearer inputs, outputs, checks, and explanations.

Evidence to Keep

Keep this page’s proof of learning as a small evidence card:

Reference Question

what you came to this appendix page to decide or clarify

Selected Rule

the rule, checklist item, or explanation you will apply

Course Link

which chapter or project this reference supports

Risk Check

treating appendix material as passive reading instead of a decision aid

Expected Output

a note that changes a route, setup, project, or review decision

Pass Check

You pass this appendix page when one vague worry has become one concrete next action tied to a chapter, project, or setup check.