| name | chapter-template |
| description | Write a new chapter for the backend_to_ml book — 10-section format, code-first style. Triggered when the user mentions "new chapter", "yangi bob", "month-XX", "new section", or asks to add content. |
Chapter template (10 sections)
backend_to_ml guides backend developers toward ML/MLOps. Every chapter follows a fixed 10-section structure. Style: code-first, theory-minimum — leave deep academic theory to Andrew Ng's course.
Language and format
- Body language: Uzbek (Latin script)
- Technical terms stay in English:
gradient descent, embedding, backpropagation, tensor, loss function
- Code examples: 10–30 lines, fully runnable (tested with
uv run)
- Emojis are rare: only for navigation headers, exercise difficulty markers, capstone badges. Not in every paragraph.
The 10 sections
# Chapter title (short, specific)
## 1. Maqsad
1–2 sentences. What and why, framed in backend-developer terms.
## 2. O'rganadigan narsalar
A 5–8 item list. "Understand X", "Write Y", "Integrate with Z".
## 3. Kerakli kutubxonalar
\`\`\`bash
uv sync --group month-XX
\`\`\`
Or the matching pyproject.toml group. Show versions.
## 4. Asosiy mavzular
### 4.1 Topic 1
Short explanation (1–2 paragraphs). Don't dive deep into theory.
### 4.2 Topic 2
...
(Usually 3–5 subtopics)
## 5. Kod misollari
\`\`\`python
# 10–30 line example
# Imports
# Logic
# Output / print
\`\`\`
Multiple examples are fine. Each one fully self-contained.
## 6. Backend integratsiya
**Mandatory section** — the book targets backend developers.
\`\`\`python
# How to use the model from a FastAPI endpoint or Django view
from fastapi import FastAPI
# ...
\`\`\`
## 7. Manbalar
- 📘 [Book name](URL) — chapter/page
- 🎥 [Course name](URL) — module
- 📰 [Article](URL)
- 🔗 [GitHub repo](URL)
3–7 references. Andrew Ng, fast.ai, Hugging Face are standard.
## 8. Mashqlar
### 🟢 Easy (1–2)
1. ...
2. ...
### 🟡 Medium (2–3)
1. ...
2. ...
3. ...
### 🔴 Hard (1–2)
1. ...
2. ...
## 9. Capstone 🏆
A mini-project that ties the chapter together. 2–5 days of work. Backend integration required.
Example: "Add ML-based fraud detection to the User model in your existing Django project."
## 10. Tekshirish ro'yxati
- [ ] I understand X
- [ ] I can write Y
- [ ] I applied Z in a real project
- [ ] I completed at least 70% of the exercises
- [ ] I finished the capstone
When creating a new chapter
- Pick the folder:
src/month-XX-*/ — existing ones are month-01-foundations, month-02-classic-ml, etc.
- Filename:
NN-topic.md (NN = two-digit ordinal)
- Add the link to
src/SUMMARY.md in the right order, before exercises.md
- Run
./scripts/sync-translations.sh — appends new msgids to po/ru.po and po/en.po
- If new terms appear, add them to
po/glossary.yaml
- Translate with
/translate ru month-XX-* and /translate en month-XX-*
Quick command
/new-chapter month-XX NN-topic — invokes this skill and delegates to the chapter-creator subagent.
