🎯 RAG FLOW — Visual Representation

🟢 1️⃣ USER QUESTION

"What is product warranty?"

⬇️

🔵 2️⃣ RETRIEVER (Vector Search Engine)

🧠 Converts question → embedding
📚 Searches vector database
🔎 Finds top relevant documents

Query → Retriever → Relevant Docs

⬇️

🟣 3️⃣ CONTEXT CREATION

📄 Retrieved documents are formatted
🧩 Combined into structured context

Context:
"Product warranty is valid for 2 years..."

⬇️

🟠 4️⃣ PROMPT BUILDING (IMPORTANT STEP)

Both are merged together 👇

-----------------------------------
Use the following context:

Context:
Product warranty is valid for 2 years...

Question:
What is product warranty?
-----------------------------------

✅ Context + Question combined
✅ Sent in ONE call to LLM

⬇️

🔴 5️⃣ LLM (Reasoning Engine)

🤖 Reads context
🧠 Understands question
✍️ Generates final answer

Answer:
The product warranty is valid for 2 years...

🌈 ULTRA SIMPLE FLOW DIAGRAM

🟢 User Question
        ↓
🔵 Retriever (Search)
        ↓
🟣 Context Created
        ↓
🟠 Context + Question
        ↓
🔴 LLM
        ↓
✅ Final Answer

💡 One-Line Summary (Interview Gold)

Retriever fetches knowledge.
LLM reasons over that knowledge.