Add Build an AI System from Scratch to Deep Learning - #985

Adds one tutorial under Python, Deep Learning.

https://github.com/Amey-Thakur/AI-ENGINEERING

Disclosure, as the guidelines ask: I am the author of this course.

Against each point in the contribution guidelines:

• Not already listed. Neither the URL nor the title appears in README.md.
• Correctly placed. Python, then Deep Learning. It is appended at the end of that section, which is in insertion order rather than alphabetical.
• Free and open. MIT licensed, hosted on GitHub and GitHub Pages. No paywall, no login wall, no newsletter signup, no email address required.
• Project-based. Fifty lessons, in which the reader writes every line. They build a classifier, a neural network with backpropagation, word embeddings, a language model, retrieval, a tool-using agent, and then an attention head, positional embeddings and a complete pre-norm transformer block. Each lesson ends with a program the reader has written and can run.
• One tutorial, one pull request.
• Format. - [Title](link), one line, no trailing whitespace, no URL shortener, the link points straight at the course.

Two things that may be worth knowing, since they are unusual for this list:

• Nothing outside the Python standard library is imported anywhere in the course. No account, no API key, no GPU, no paid service, and the lessons run with the network closed. A program in the repository checks that on every change.
• Every number shown on a lesson page is compared against what the code actually prints, on Linux, macOS and Windows, on every push and again nightly, so the pages cannot drift from the code.

I was not able to run scripts/check_readme.py lint in my environment. I checked the rules it enforces by hand: correct format, no trailing whitespace, no duplicate or shortened URL, and no table of contents change needed since no new section was added. Please let CI be the judge, and I will fix anything it finds.