AI CareerPathJob board
← All projects
📚

Production RAG Chatbot

Intermediate

5–7 days · 7 milestones

Build a chatbot that answers questions from your own documents using RAG — chunking, embeddings, vector search, and LLM generation with source citations.

Difficulty

Intermediate

Duration

5–7 days

Milestones

7 steps

Interview Qs

4 questions

🛠 Tech stack

PythonLangChain or LlamaIndexQdrant or PineconeOpenAI EmbeddingsFastAPINext.js or Streamlit

✅ Prerequisites

  • Python (intermediate)
  • Basic ML concepts
  • REST APIs

Step-by-step milestones

1

Document ingestion pipeline

Load, clean, and chunk documents.

2

Generate embeddings

Embed chunks using OpenAI or sentence-transformers.

3

Set up vector DB

Store and query embeddings in Qdrant or Pinecone.

4

Build retrieval + generation

Retrieve top-K chunks and feed to LLM.

5

Add source citations

Return source document name and page with each answer.

6

Evaluate quality

Measure retrieval precision and answer faithfulness.

7

Deploy

Containerize with Docker and deploy to cloud.

Skills you'll build

RAGEmbeddingsVector DBLLM APIChunking strategies

📝 Interview questions you'll face

  • 1.What chunking strategy did you use and why?
  • 2.How do you handle questions outside your knowledge base?
  • 3.How do you evaluate RAG pipeline quality?
  • 4.What is the difference between dense and sparse retrieval?

🌟 Portfolio guidance

Build it on a real corpus (company docs, research papers). Write a blog post explaining your chunking + retrieval choices.