AI Model Design Workshop with Kyle Alexander
There are hundreds, if not thousands, of courses that teach you how to use AI tools, but almost none that teach you about model design. At its core, model design involves shaping AI behavior, deciding what a model should and shouldn’t do, and determining the rules it’ll live by.
It’s an emerging field with very few practitioners, so the people who can do this work are valuable in a way that’s hard to overstate.
Why does this matter now?
A lot of conversations about AI are positioning speed and output as virtues. Model design is the counter-practice; the slower, more deliberate, and far more durable approach.
What you'll leave with
By the end, you'll have a working vocabulary for model design and a framework for evaluating AI outputs against stated intent. More practically, you'll have language that makes your work legible to PMs, engineers, and executives who are trying to figure out what designers actually do in an AI-first org. That legibility matters when you're making the case for your scope, your title, or a seat in earlier product conversations.
What this is
A working session for designers who want to think at the level of model behavior. They’ll leave with vocabulary, a framework, and a point of view they can use in product conversations.
What this is NOT
A course on AI tools, a prompting tutorial, a prediction about where the industry is going, or a proclamation that design is dead. This is about the design decisions that sit upstream of all of that.
Topics covered
The foundations
What model design is and isn't
How it differs from prompt engineering
The difference between model behavior and model output
What "alignment" actually means in practice vs. in research
How models get their personality, tone, and constraints
System prompts: what they are, what they do, what they can't do
Behavioral design
Writing behavioral guidelines
Defining what a model should and shouldn't do
Hard rules vs. soft defaults
How to think about edge cases and failure modes
Consistency as a design material
Evaluation
What evals are and why they matter
How to evaluate outputs against stated intent
Benchmarks: what they measure and what they miss
The gap between benchmark performance and real-world behavior
Qualitative vs. quantitative evaluation
Red-teaming basics for designers
How to know when a model is working
Content and product design overlap
Where content design skills transfer directly
Voice and tone as model design inputs
How UX writing decisions become model decisions
Reading product copy through the lens of model behavior
Designing for variable-length, non-deterministic outputs
Organizational layer
How to make the case for model design inside a company
Why most orgs don't have language for this yet and how to give it to them
Positioning content and product designers as decision-makers
How model design decisions affect product direction
The relationship between model design and product narrative
How to read a competitor's AI product through its behavior
About the instructor
Kyle Alexander has spent a decade convincing companies that content design matters and then proving it. He's done it at Fidelity, Credit Karma, and Square.
His current focus is a response to the vibe coding era, positioning content and product designers as the people who decide how and why AI communicates. He teaches this stuff because not enough people do.
Catering
Catering will be from BigNorm (Korean) and feature bibimbap, vegetarian dumplings, spicy tuna kimbaps and more.
Schedule
6:00 PM - Doors open + mingling
6:15 PM - Workshop starts
8:45 PM - Wrap up
9:00 PM - Doors close
Allergies & dietary requirements
While many of the options are gluten free, the prevalence of soy sauce in Asian cuisine means that we can not guarantee a contact-less kitchen for folks with Celiac disease.
If your allergy is severe, please contact us to work out accommodations.
🐝 Community Guidelines 🐝
Designer Friends Community Guidelines & code of conduct