Serving schools, colleges & students across India Sign inRegister freeFAQsHelpcontact@ghosla.io
GHOSLAGrowth Hub for Opportunities, Skills, Learning & AchievementGrow•Learn•Achieve Login Get Started

Solutions (For Institutions)

School ERPComplete school management College ERPComplete college management University ERPUniversity-wide management Coaching Classes ERPCoaching institute management

For Students

Ghosla AIAI-powered education & career assistant AI Career TestFind the careers that fit you Education LoansPlan, apply & track your loan Skills & CertificationsFree courses, Skill Score, certificates All 103 CoursesJob-ready courses, free right now Find Better SchoolsDiscover the right school for you Find Better CollegesDiscover the right college for you

Discover

Browse by City66 cities across India

Generative AI with Python: RAG Apps & AI Agents

Build retrieval-augmented chatbots and tool-using AI agents in Python, then evaluate and ship them safely

Ghosla Academy · AI Team

Most useful generative AI apps are not just a chat box. They search your own documents before answering (retrieval-augmented generation, or RAG) or they take actions by calling tools (AI agents). This course teaches both from first principles in Python, starting with how large language models work and how to call one safely, then moving to embeddings, chunking, vector search and a complete RAG chatbot with citations. In the second half you build agents that call Python tools through JSON schemas, add guardrails such as step limits and human approval, and learn to evaluate quality with a small test set instead of guessing. Everything runs in Google Colab or on a basic laptop using free tiers or a local model. Explanations are in simple English with occasional Hinglish.

What you will be able to do

  • Explain tokens, context windows, temperature and why LLMs can hallucinate
  • Call an LLM API from Python safely with keys in environment variables and structured JSON output
  • Create embeddings, chunk documents and run similarity search with NumPy and ChromaDB
  • Build an end-to-end RAG pipeline that cites its sources and refuses when the answer is not in the documents
  • Design an AI agent loop with tool calling, JSON schemas, step limits and human approval for risky actions
  • Evaluate RAG and agent quality with a test set and share a simple Streamlit demo

What you build

  • RAG chatbot that answers questions from a college prospectus or rulebook with citations
  • Personal expense assistant agent that uses Python tools over a UPI transactions CSV

What you need to start

  • Comfortable Python: functions, dictionaries, lists and installing packages with pip
  • A laptop with a browser and a free Google account for Colab; a free-tier LLM API key or a local model via Ollama

Curriculum · 5 sections, 20 lectures

How LLMs Work4 lectures
  1. LLMs in Plain Language: Tokens, Context and ProbabilitiesFree preview13 min
  2. Calling an LLM API from Python SafelyFree preview14 min
  3. Prompt Patterns for Developers: System Prompts, Few-Shot and JSON Output12 min
  4. Quiz: How LLMs Work6 min
Embeddings and Vector Search4 lectures
  1. Embeddings and Cosine Similarity14 min
  2. Chunking Documents the Right Way12 min
  3. Vector Databases with ChromaDB13 min
  4. Quiz: Embeddings and Vector Search6 min
Building a RAG App4 lectures
  1. The RAG Pipeline End to End15 min
  2. Grounding, Citations and Saying "I Don't Know"11 min
  3. Project: RAG Chatbot for a College Prospectus or Rulebook20 min
  4. Quiz: Building a RAG App7 min
AI Agents and Tool Calling4 lectures
  1. What Makes an Agent: Loop, Tools and Memory14 min
  2. Tool Calling with JSON Schemas14 min
  3. Agent Safety: Permissions, Limits and Human Approval11 min
  4. Quiz: AI Agents and Tool Calling7 min
Evaluating and Shipping4 lectures
  1. Evaluating RAG and Agents13 min
  2. Building a Simple Streamlit UI and Sharing It12 min
  3. Project: Personal Expense Assistant Agent with Tools20 min
  4. Quiz: Evaluating and Shipping7 min

Who this is for

  • Python learners and freshers who want to build real generative AI projects for their portfolio
  • Developers moving from simple chatbot prompts to production-style RAG and agent apps
  • Final-year engineering and BCA/MCA students looking for a strong AI project