[I wrote on this topic in 2024 as well as in 2025. All the references shared in those two posts are still valid]
Two paths for LLM
As described in the famous “The Large Language Model Course”, when it comes to contributing to the world of LLMs, there are two paths:
- LLM Science: This is about building Large Language Models (LLMs).
- LLM (AI) Engineering: This is about building applications using LLMs
This post focuses on resources useful for the second track, LLM Engineering.
Start by using
My first piece of advice to anyone interested in getting started with GenAI is to subscribe to one of the paid services: ChatGPT Plus, Claude or Gemini and then actively use it in daily life (both at home & work). The $20/month might feel like an expense, but for me, it has been the best personal investment I have been making for the past two and half years.
This detailed blog by Ethan Mollick compares the current available options and helps to decide the best way to spend that $20.
Learn from the master
Next, I will recommend these two videos by Andrej Karpathy:
Learn by building
The best way to learn about LLM Engineering is by building. Depending on your preference, you can chose to learn either by writing code (for a use case of your choice) or by working on courses which are focused on building.
Courses
Here are two courses from deeplearning.ai’s which can help someone to get started (in sequence) with the fundamentals: using LLM through APIs, building RAG as well as Agents. These courses reduce the friction of learning by providing online environment to write code and use LLMs over API.
- Retrieval Augmented Generations: Free to Audit (Multiple Days to Weeks)
- Agentic AI: Free to Audit (Multiple Days to Weeks)
Code heavy tutorials
Cookbook
OpenAI’s cookbooks are code heavy and allows someone to get started with almost any use cases. However, you need to set up your environment (Python, API Key etc.) to write/execute code. This is one of my favorite go to place.
I prefer the above over langchain’s tutorial.
Learn from a book
If you want to learn by reading, you can follow Chip Huyen’s AI Engineering.
arnab's blog 