What the plan looks like
6 weeks · 4 sessions a week, about 45 min each
Where you end up: You can explain how AI agents work, write simple Python scripts, and call an LLM API to build a tiny tool-using agent.
Phases
- AI Foundations · week 1
- Python Basics · weeks 2 to 3
- Talking to an LLM · week 4
- First AI Agent · weeks 5 to 6
Week 1AI Foundations
- Day 1Learn what an LLM is
Watch the first two lessons of AI for Everyone before noon. Write 3 bullet notes on what an LLM can and cannot do.
- Day 3Understand AI agents
Read one beginner article on AI agents and write a one-line definition plus the four parts: model, tools, memory, loop.
- Day 5Practice prompting
Spend 20 minutes in ChatGPT or Gemini testing 5 prompts on a CA topic and note which prompt style gave the best answer.
- Day 6Map voice agents and automation
Watch one overview each of voice agents and n8n automation, then write 2 business use cases for a CA firm.
Week 2Python Basics
- Day 8Set up Google Colab
Open Google Colab and run print('Hello') in a new notebook. Each Python session counts as done when your notebook cell runs without errors.
- Day 10Variables and data types
Complete the variables and strings lessons and write 5 lines storing a client name, fee and GST rate.
- Day 12If statements
Write a script that prints whether an income falls above or below a tax slab threshold.
- Day 13Lists and loops
Loop over a list of 5 invoice amounts and print their total.
Week 3Python Basics
- Day 15Dictionaries
Store 3 clients in a dictionary with name and balance, then print each one with a loop.
- Day 17Functions
Write a function calculate_gst(amount, rate) and call it 3 times with different values.
- Day 19Libraries and pip
Install the requests library in Colab and fetch data from one free public API.
- Day 20Python review
From memory, without notes, rewrite the GST function and invoice loop in a fresh notebook.
Week 4Talking to an LLM
- Day 22Get a Gemini API key
Create a free key in Google AI Studio and store it in Colab secrets. Done means the key is saved and never pasted into shared code.
- Day 24First API call
Send one prompt to Gemini from Python and print the reply.
- Day 26System prompts
Give the model a role as a CA study tutor and compare 3 answers with and without it.
- Day 27Build a chat loop
Write a while loop that keeps chatting until you type quit, keeping message history.
Week 5First AI Agent
- Day 29Start Hugging Face Agents course
Complete Unit 1 reading on thought, action, observation. Write the agent loop in your own words in 4 lines.
- Day 31Learn function calling
Read the Gemini function calling guide and note how a tool is described to the model.
- Day 33Build a GST tool agent
Give the model your calculate_gst function as a tool and ask it 3 GST questions.
- Day 34Add a second tool
Add a tool that returns a client balance from your dictionary and test that the agent picks the right tool.
Week 6First AI Agent
- Day 36Try an agent framework
Run the smolagents starter example from the course and compare it with your hand-built agent in 3 notes.
- Day 38Explain and plan next
Write a half-page summary of how your agent works and pick voice agents or n8n automation as your next chapter.
Days count from the day you save the plan. Once it's yours, Targela fits the sessions around your own week and pace, and you can change any step.