AI Engineering Core Track – LLM Engineering – RESOURCES

, , , , , , , , , ,

Here are links to all my courses including this one, with coupons

I absolutely loved making this 8 week course on AI Engineering – here is my full Udemy curriculum with coupons. We build a number of chunky commercial projects. The final project is a fitting conclusion — we build an Agentic AI solution that solves a complex commercial problem. It performs way better than I imagined possible.

2026 Update

As requested by many students – I’ve rolled out freshly updated videos! They are a swap-in replacement for the original videos.

I generally try to avoid changing videos often because it can be quite disruptive to everyone. I update the code frequently, but aim to keep the videos constant. But we’ve reached a point where the benefit of new content outweighs the downside.

What’s changed in the new version?

I’ve switched to using Cursor and uv, instead of Jupyter Lab and Anaconda. Much easier environment setup. I’ve also used newer models (like GPT-5 and 5.2, Claude Sonnet 4.5, Gemini 3), and I’ve added in tons of new tricks and techniques along the way. But the overall learning journey is the same.

I also changed the name of the course from “LLM Engineering” to this new name, to improve clarity.

Oh and also, I’m a year older. Sadly no wiser. 😁

Repo, Setup and Slides
  • The Github repo for the course
  • The README with setup instructions
  • All the slides for the course
  • A link to this course on Udemy with my companion courses and how they fit together – if you’re interested in going deep on Autonomous Agents, you might consider this as a follow-up course – the projects are equally surprising and amazing! And the Production course takes everything to the next level.
  • A useful FAQ for common questions
  • The Proficient AI Engineer directory for graduates of my courses

And if you wish: please connect with me on LinkedIn, follow me on X, subscribe to me on YouTube and register below! All the multi-modal forms of me 😂

The definitive answer to the most common first question!

People from a non-Data Science background often ask me a great question during Week 1: so what exactly are these “parameters” that we keep hearing about?? I’ve made this short video to explain what they are, and how they give GPT its super-powers, followed by a peak inside GPT.

Important note: updating your code after each week

I regularly push updates to the labs, including more tips, business applications and exercises. It’s worth bringing in the latest code at the start of each week, beginning with Week 2.

First, from the llm_engineering project root directory, pull in the latest code from git and merge in any of your changes. Instructions here for those less familiar with git.

Then update your environment to bring in the latest libraries by running uv sync

Contributing to the repo

Many students have contributed their own solutions and extensions to the repo. I’m incredibly grateful! I love seeing your progress and innovative ideas, and it adds value for everyone else on the course. As an added benefit, you get recognition in GitHub as a contributor to the repo.

If you’re interested in adding your work, please submit a Pull Request and I’ll merge as soon as possible. Instructions are in the Git guide in the guides folder and explained more here.

Another fun example project

Here’s a video extra on a project to have LLMs compete that shows how easy it is to use different Frontier APIs, and the benefits of writing your own lightweight LLM abstraction.

Frontier models – Web interface
  1. ChatGPT (latest model GPT-4o and o1) from OpenAI
  2. Claude (latest model Claude 3.5 Sonnet) from Anthropic
  3. Gemini Advance (latest model Gemini 2.0 Flash) from Google
  4. DeepSeek (latest models DeepSeek R1 and V3) from DeepSeek AI
  5. Le Chat from French AI powerhouse Mistral
  6. Chat with Command R+ from Cohere
  7. Meta.ai (model is Llama 3) from Meta
  8. Perplexity (latest model is Perplexity Pro) from Perplexity.ai

Here’s a review of OpenAI’s chat model from earlier this year, GPT-4.5:

Frontier models – API
  1. GPT API from OpenAI
  2. Claude API from Anthropic
  3. Gemini API from Google
  4. DeepSeek API from DeepSeek AI
Other useful links on models

The seminal 2017 paper ‘Attention Is All You Need’ from Google scientists that brought about the Transformer is here. This sentence from the Abstract says it all:

We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely.

The famous paper ‘On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?’ that discussed bias and deception is here.

The prompt generator from Anthropic is described and linked here.

And here’s the Vellum leaderboards including costs and context windows.

The paper on the Chinchilla Scaling Law, describing how the scaling of model parameters is proportional to the size of your training data, can be found here.

Here is the game I made, Outsmart, that pits models against each other in a battle of negotiation.

Common tools used in LLM engineering:
  1. Hugging Face – the go-to hub for models, datasets, leaderboards and even applications, and the authors of many essential open source frameworks including the pioneering transformers library
  2. LangChain – open source framework that provides abstractions connecting multiple LLM operations under a simple API
  3. Gradio – a ridiculously simple UI framework that lets you create prototype UIs in one line of code, no frontend experience needed
  4. Weights & Biases – tooling to analyze and visualize during training
  5. Google Colab – write, evaluate and share notebooks remotely on a box in the Google Cloud
  6. Modal.com the serverless AI platform
Not covered in this course: using a Managed Service
  1. Amazon Bedrock is the managed service from AWS:
    “The easiest way to build and scale generative AI applications with foundation models”
  2. Vertex AI is the managed service from Google Cloud:
    “Innovate faster with enterprise-ready AI, enhanced by Gemini models”
  3. Azure Machine Learning is the managed service from Microsoft.
    “Build business-critical ML models at scale”
Links to the Google Colabs

You should be able to use the free tier or minimal spend to complete all the projects in the class. I personally signed up for Colab Pro+ and I’m loving it – but it’s not required.

Learn about Google Colab and set up a Google account (if you don’t already have one) here

Then the links to the colabs themselves are in the repo for the respective day. So, in the week3 folder, the individual day notebooks contain the link to the right colab to use.

The Leaderboards and Arenas
Real-world examples of LLMs making commercial impact
The Extra Extra Project for Fun

I mentioned my experiment to train an LLM on my 240,000 text message history. My write-up of the journey is here, and the subsequent blog posts take you through the adventure of replicating this yourself!

Finally

Somehow you made it all the way to the end of the resources — thank you! If you’re not completely fed up with me by now, then please connect with me on LinkedIn! I’d love to stay in touch and I’m always available if you have feedback, questions or ideas.

Leave a Reply

Discover more from Edward Donner

Subscribe now to keep reading and get access to the full archive.

Continue reading