My 7-Day AI Learning Journey: Starting with LLMs, RAG and AI Agents

I've been wanting to learn AI for a while, but honestly, getting started was the hardest part.

There are so many topics—LLMs, RAG, AI agents, automation, APIs, embeddings, vector databases, machine learning, fine-tuning—and it was difficult to figure out where to start and what to learn first.

So instead of jumping between different videos and courses, I've decided to take a different approach.

For the next seven days, I'll learn the fundamentals of LLMs, RAG, AI automation and AI agents, while working on my own project alongside the learning.

Why Hamara Sheher?

The project I'll be using as my practical learning ground is Hamara Sheher.

Hamara Sheher started as an idea to build a modern platform for discovering cities, starting with Hamara Raipur.

Now, I want to take it beyond just a website.

As I learn AI, I'll try to apply what I learn directly to the project.

Instead of building small practice projects that I may never use again, I want Hamara Sheher to become my practical resource, testing ground, and eventually my final portfolio project.

My 7-Day AI Learning Plan

Day 1 — LLMs

I'll start by understanding Large Language Models (LLMs) and how AI tools like ChatGPT work.

I'll learn about prompts, tokens, context, APIs, structured responses, and hallucinations.

Then I'll start using an LLM with Python.

Day 2 — Embeddings & Vector Databases

Next, I'll learn how AI can understand the meaning of information.

I'll explore embeddings, semantic search, and vector databases.

This will help me understand how AI can find relevant information from large amounts of content.

Day 3 — RAG

Then I'll move to RAG (Retrieval-Augmented Generation).

I'll learn how an AI system can use external information before answering a question.

I'll build a simple system that can chat with documents and use this understanding later in Hamara Sheher.

Day 4 — Advanced RAG

I'll then explore how RAG can become more accurate and useful.

I'll learn about better retrieval, search, sources, and reducing incorrect answers.

Day 5 — AI Tools & Automation

Next, I'll learn how AI can do more than just answer questions.

I'll explore how AI can use tools, functions and APIs to perform actions and automate tasks.

Day 6 — AI Agents

Then I'll move into AI agents.

I'll learn how an AI system can decide what it needs to do, use different tools, and complete a task.

Day 7 — Training & Fine-Tuning

Finally, I'll try to understand what training an AI model actually means.

I'll explore the difference between prompting, RAG, fine-tuning, and training.

Learning Without Videos

For now, I've decided to learn without relying on videos.

I'll use documentation, articles, examples, experiments, and most importantly, building.

The idea is simple:

Learn something → apply it to a project → understand what went wrong → improve it.

I'll also continue working on Hamara Sheher in parallel.

This way, I'm not just learning what RAG or AI agents are.

I'm learning why I would actually use them.

What I Want to Build

By the end of this journey, I want to have a better understanding of:

LLMs → RAG → AI Tools → Automation → AI Agents

And I want to gradually bring these technologies into Hamara Sheher.

Eventually, I want Hamara Sheher to become more than a website.

I want it to become a project that demonstrates what I've learned about AI, automation, marketing, product development, and technology.

I don't know exactly where this journey will take me yet.

And that's actually the point.

This is Day 1. I'm learning, building, experimenting, and documenting the journey as I go.