I’m so excited to be running another LLM Workshop for AI engineers ready to roll up sleeves and get coding! Here are some essential links to get started:

1. The class repo
2. I also have intensive online courses that cover this material (and tons more) in much detail! This covers my training curriculum on LLM engineering and AI Agents.
3. I’m building a directory of Proficient AI Engineers who complete all my courses
4. My related O’Reilly events on LLMs

The role of LLM Engineer is the new, emerging profession at the crossroads of Software Engineering, Data Science and ML Engineering. With Sam Altman’s holiday blog post predicting Agents may enter the workforce this year, there’s never been a better time to be building Agents! And that is exactly what we will do in the LLM workshop.

In just a few hours, we’ll build a full Agentic AI platform from scratch. Same commercial idea as my prior events, but taking it to a whole new level. These agents will be fully autonomous — self-directing, making 23 separate LLM calls per decision, and perhaps occasionally going rogue. (What’s a live event without a little drama..)

Here are some more resources to accompany the session.

The background to LLMs

This is one of the most common questions that I’m asked. “Can you explain how LLMs actually work, for someone new to it?” After many attempts, I’ve come up with this set of videos, and I think I’ve finally nailed it! Let me know if you agree!

Segment 1: LLM Engineering

BACKGROUND PAPERS

• Turing’s seminal 1950s paper that introduced the Turing Test
• the 2017 paper “Attention is all you need” that invented the Transformer

WEB UI CHAT INTERFACES FOR FRONTIER MODELS

1. ChatGPT (latest models GPT-4o and o1) from OpenAI
2. Claude (latest model Claude 3.5 Sonnet) from Anthropic
3. Gemini Advance (latest model Gemini 1.5 Pro) from Google
4. Chat with Command R+ from Cohere
5. Meta.ai (model is Llama 3.1) from Meta
6. Perplexity (latest model is Perplexity Pro) from Perplexity.ai

API INTERFACES

1. GPT API from OpenAI
2. Claude API from Anthropic
3. Gemini API from Google

I mention the Vellum leaderboard, which has a very useful table of API costs about half way up.

Segment 2: Models, Tools and Techniques

It’s important to read Anthropic’s post on building effective agents – it resonates strongly with me.

For more on LLM leaderboards and benchmarks, please see the resources for my Choosing the Right LLM event.

Tools of the trade:

1. PyTorch
2. HuggingFace
3. LangChain
4. Gradio
5. Weights & Biases
6. Google Colab

Segment 3: Getting Hands-On and Wrap

If you made it to the final section – I’m so grateful – and please connect with me on LinkedIn here! I love building up an AI Data Scientist community and I welcome connections. Message me if you have any specific questions or if you’d like to sound me out.

If you want to take your learning to the next level, you could consider my 8 week immersive course on mastering LLM engineering. By the end, you’ll be able to build your own Agentic LLM solutions that outperform frontier models.

The final treat at the end

You’ll find my code and write-up for simulating myself with my 240,000 text message history here. I’d love to hear if you try this for yourself!

Finally, here’s an entertaining LLM game I wrote a few months ago that has LLMs battle against each other, with a write-up and the code.

That’s it! I hope you’ll have the time to come back and try out the code for yourself — enjoy the journey, and please do message me if I can help.