Kevin CaoProduct ManagerAI-ForwardThought-LeaderAlways LearningCoffee Enthusiast
ABOUT THE BUILDER

Hi, I’m Kevin

I’m a product manager and builder who loves turning ambiguous ideas into products people actually use.

I’ve spent my career launching 0→1 products, scaling experiences to hundreds of thousands of users, and solving customer problems across a wide range of industries and user groups. My work spans product strategy, customer research, rapid prototyping, experimentation, AI-powered experiences, and taking products from an early idea through launch and iteration.

I’m particularly interested in how AI and personalization can make products more useful without making them more complicated. I enjoy getting close to the problem, building something quickly, putting it in front of users, and using what I learn to make the next version better.

THE PROBLEM

Finding great coffee shouldn’t require knowing everything about coffee

Generic ratings tell you what everyone likes. Specialty coffee tools often assume you already know exactly what you want. I wanted to explore a simpler experience for people who are still learning their own taste: start with what you actually drink, capture what you liked or disliked, and let those signals make the next recommendation more useful.

01

Something to drink

Get a café drink matched to your history and what you’re in the mood for right now.

02

Something to brew

Explore beans and home-brew ideas informed by the flavors and methods you’ve enjoyed.

03

Somewhere to go

Discover cafés whose drinks fit your preferences — or ask for a surprise.

THE PRODUCT

Coffee recommendations that get better as they get to know you

Excellatte is an AI-powered coffee discovery experience that learns from what you drink, what you rate, and what you say — then uses those signals to recommend what you might enjoy next.

Hand-drawn steaming coffee beside a rustic bag of coffee beans
HOW I SOLVED IT

Your coffee history becomes your recommendation engine

Ratings, written notes, drink history, and recommendation feedback become signals that gradually build a more useful picture of what you enjoy. Recommendations stay flexible: reject one, explain why, change your current mood, or deliberately venture outside your comfort zone.

DrinkLog what you tried
→
RateTell us what you thought
→
LearnBuild your taste profile
→
DiscoverGet a better match
→
RepeatImprove every time

AI will become a core part of how great product managers work

I believe learning to work effectively with AI will make me a better product manager — helping me explore ideas faster, build more independently, and spend more time on the judgment that matters. As these tools become part of everyday product development, I wanted to start building that relationship now rather than wait for the workflow to become standard.

Excellatte gave me a practical way to learn by building

The product started with a simple question: Could your coffee history help predict what you’ll enjoy next? Building Excellatte gave me a practical way to test that idea while learning where AI genuinely accelerates product work, where it needs direction, and where human judgment still matters most.

0 → 1 Product DevelopmentTurning an ambiguous concept into a scoped, interactive MVP.
AI & PersonalizationUsing ratings, language signals, feedback, and metadata to shape recommendations.
Search & RecommendationsExploring ranking, similarity, cold start, feedback loops, and serendipity.
Rapid PrototypingUsing AI-assisted development to test assumptions and move quickly from idea to interaction.
Product JudgmentConsciously limiting scope to learn before investing in scale or infrastructure.
Customer-Driven IterationCritiquing each prototype, incorporating feedback, and improving the next version.
TRY IT YOURSELF

What are you in the mood for?

Try Excellatte using a sample coffee profile and see how recommendations adapt to your choices.