Course description

From your first chatbot to your own AI agent, deployed online

A chatbot can only talk. An AI agent can act: it checks your deadlines, searches real sources, books a study room and asks you before anything risky. This course shows you how agents work and how to build them in Python, one clear step at a time.

Every video follows UniBuddy, a personal agent for a university student, and adds one new skill to it. You see the idea in animations, read the real code, watch it run, learn the common mistakes, and finish with a quiz and practice tasks. You can use free models (Gemini or Groq), OpenAI, or an offline model on your own laptop with Ollama and Gemma.

What's inside

1. What Is an AI Agent? The agent loop, chatbot vs workflow vs agent, and the five building blocks.
2. Building Your First Chatbot: API keys, free and offline models, roles, a system prompt and a web chat.
3. Giving Tools to Your Chatbot: function calling and the tool loop: GPA, deadlines and live weather.
4. Memory, Short-term and Long-term: the context window, summaries, saved facts and embeddings.
5. Planning and the ReAct Pattern: Thought, Action, Observation, plan-and-execute and stopping rules.
6. A Complete Research Assistant Agent: Wikipedia and arXiv tools, and a report with citations.
7. Multi-Agent Systems: a Planner, Researcher, Writer and Reviewer working as a team.
8. Chatbots and Agents Together: a router, answers from a handbook (RAG), confirmations and human handoff.
9. Safety and Best Practices: prompt injection, least privilege, limits, logging and evaluation.
10. Final Project and Deployment: everything in one agent app, tested and deployed on Streamlit Cloud.

Projects you build

• Study Buddy chatbot, in the terminal and on the web
• A tool-using agent that calculates a GPA, counts days to exams and checks the weather
• A Research Assistant that searches real sources and writes a cited report
• A Group Project Helper made of four agents
• UniBuddy, your own personal agent, live on the internet

Course features

• 10 videos, about 2.5 hours in total
• Complete, runnable Python code for every video
• Works with free models (Gemini, Groq), OpenAI, or offline with Ollama and Gemma
• Animated explanations, live demos and a 4-question quiz in every video
• Practice tasks in every video and a final project with a rubric

Who this course is for

• University students from any department who know basic Python
• Beginners in AI who want to build something real, not just use ChatGPT
• Anyone curious about how AI agents work under the hood

The lessons are written and taught by Touqeer Abbas. The on-screen presenter is an animated avatar of the instructor, and the voice is synthetic.

What will i learn?

• Explain what makes an AI agent different from a chatbot and a fixed workflow
• Call a language model from Python with Gemini, Groq, OpenAI or an offline model
• Build a chatbot with a persona, a conversation history and a web interface
• Give a model tools with function calling, and write the tool loop
• Add short-term and long-term memory to an agent
• Make an agent plan and act step by step with the ReAct pattern
• Build a research agent that searches real sources and cites them
• Design a team of agents, and a hybrid chatbot and agent system with a router
• Protect agents from prompt injection with confirmations, limits, logging and tests
• Deploy your own AI agent as a web app on Streamlit Community Cloud

Frequently asked question