How AI Image Models See the World: A Practice Run

Are you curious how AI image models actually learn, and how understanding that process can help you use them more creatively?
Hey, I’m Walt, an AI engineer working across generative AI, visual development, and model training.
I have a master’s-level lecture coming up, and I’d like to use this opportunity to do a free practice session on the 28th — part rehearsal, part open session, part chance to meet new people and learn something new together.
In this session, I’ll be combining the last three years of my experience into an accessible introduction to how AI image models are trained, how they “see” patterns, and how that understanding can help you push the tools further in your own work.
A big part of the session is about showing that generative AI does not replace human creativity. The more you understand how the model thinks, the better you can direct it, shape it, and bring your own artistic judgment and flair into the process.
Some of the key areas I’ll be going through include:
how AI image models work in simple terms
diffusion and what the model is actually doing
why custom training is needed
datasets, captioning, and LoRA training
consistency, iteration, and troubleshooting
how artistic direction and decision-making shape the final result
practical use cases from training to production workflow
I’ll also be sharing examples from my own work, including case studies, production workflows, and storyboarding examples.
This is a free informal session, so questions, feedback, and honest reactions are all welcome.
Come along if you’re interested in generative AI, want to understand model training in practice, or just want to listen to a lecture.
Please note: no photos of the presentation, much thanks.
My number +44 7477653591 . Contact me if you need access to the building.
@Enclave Bar Lounge