MY_PRODUCT.md

My Product: [Your AI Feature] for [Your Product]

This document tracks what we're building across all 25 sessions.

Started: Main model: Claude Opus 4.5

---

The Product

[Your Product Name] - [Brief description of your product]

Users: [Who uses your product]

---

The AI Feature: [Feature Name]

[Brief description of the AI feature you want to build]

Feature Vision (by end of course)

User: "[Example request]"

Your AI Feature:
1. [Step 1 - what it does]
2. [Step 2 - what it does]
3. [Step 3 - what it does]
...

---

How Each Session Builds the Feature

Phase 1: Foundations

Session	What We Add	Your Feature Capability
1	LLM basics
2	Prompt patterns
3	Embeddings
4	Vector DB
5	Chunking
6	RAG

Phase 2: Agentic AI

Session	What We Add	Your Feature Capability
7	Tool use
8	Extended Thinking + ReAct
9	Memory
10	Advanced RAG
11	Corrective RAG
12	Agentic Search (RAG alternatives)
13	Multimodal

Phase 3: Multi-Agent Systems

Session	What We Add	Your Feature Capability
14	Multi-agent patterns
15	CrewAI
16	LangGraph
17	OpenAI SDK
18	Google ADK
19	MCP

Phase 4: Production

Session	What We Add	Your Feature Capability
20	Evaluation
21	Safety
22	RLHF
23	System design
24	Fine-tune vs RAG vs Agentic
25	Capstone

---

Session Notes

Session 1: LLM Fundamentals

What we learned:

Feature application:

---

(Notes for each session will be added as you progress)

---

Key Questions to Answer By Course End

1. Build vs buy: Build custom or use existing tools?
2. RAG vs agentic vs fine-tune: Which approach for which data?
3. Quality vs speed: How much reasoning is worth the latency?
4. Safety vs utility: How strict should guardrails be?
5. Cost model: How will you price/budget this feature?

---

Final Deliverable (Session 25)

A complete product spec for your AI feature including:

• Feature requirements
• System architecture
• Build vs buy decisions
• Success metrics
• Rollout roadmap