Talk in English - US at PHP Tek 2026
Short URL: https://joind.in/talk/6509b
(QR-Code (opens in new window))
Dive into the world of LLMs and how to get the most out of them. In this session you will learn about prompting techniques, agent patterns, how to mitigate hallucination and much more. This talk teaches you how LLMs work under the hood and why they do what they do so you can understand why the patterns and techniques work. This will use Node.js as the primary language of example but techniques can be applied in all other languages.
Comments
Please login to leave a comment
This was way over my head. It assumed I'm already familiar with all the terminology of agents, how to nest them, etc. I'm not, so without concrete examples I have no way to translate this advice into actual things I would so.
[COULD] Provide High-Level Context: Start with a clear, high-level overview of what you are describing and what the final result will look like before diving into the individual components. This will help the audience follow the structure and prevent the talk from feeling a bit disjointed. It felt with like a bunch of random “things to know” without a common goal.
[SHOULD] Optimize Text Size: Increase the font size on the slides. There is currently too much empty white space, and larger text will improve readability.
[MUST] Bring Candy: Ensure candy is provided for the audience.
Talk information was good. Slides were really hard to read - either make the text bigger, or link the slide here before the talk. Presentation was fun and loose, which was good for end of the conference. I was able to dabble with local LLMs during the talk, applying what was being discussed and tinker with it.