AI Lab

A 6-Phase
AI Design Process

A practical, end-to-end workflow for taking ideas from exploration to a working build.

- keep it vague so we have more wiggle room, so scope stays flexible
- first ideate, then make visuals, so we don’t jump to solutions
- they wanna promote in NY, so the objective should reflect that
- give me visual options, so I can compare directions before choosing
- show that portion well… or remove it, so weak content doesn’t stay by default
- give me alternatives and references, so I’m not locked into the first idea
- make a prompt for Codex to implement this, so I don’t miss gaps in accuracy
- replace only the rooftop model, so the rest stays stable
- turn this into something testable, so I can judge the interaction properly
- why is browser memory 827 MB, so I know what the implementation is really doing
- this table looks kinda unprofessional, so polish doesn’t get ignored
- check accuracy, UX and feasibility, so output isn’t accepted at face value
- don’t affect any other screens, so fixes stay contained
- modify only requested files, so unrelated work stays untouched
- just focus on this one, so the iteration stays precise
- don’t invent content, so the final output stays evidence-based
- give it one final check before I send it, so errors don’t reach stakeholders
- preserve the existing implementation, so working behavior isn’t lost

02

Projects

AI experiments and prototypes.

Mindle

A weekly UX word game designed to turn a short puzzle into learning, reflection, and community participation.

View case study

Space Shooter

My first AI-assisted build: from a rough game idea in ChatGPT to a playable experience in Cursor.

View case study

Perspective

A Workshop 6 concept game that makes AI-assisted decision-making something people can feel.

View case study

03

Workshops

AI in Design Workshop Series

Workshop

AI for Designers

A grounding session for designers to understand what AI is, what it can and cannot do, and how to approach AI tools more intentionally instead of relying on trial and error.

My Role: Co-designed and co-hosted
Audience: Designers, product teams, and curious builders
Format: 90-minute interactive virtual workshop · 68 joined

AI for Designers workshop visual

Key takeaways

Understand AI fundamentals and where its limits are.
Build an intentional workflow instead of copying prompts.
Explore how AI can support everyday design work.

What I taught

My AI workflow across design and build, including a live demo from a ChatGPT idea to a working interaction in Cursor.

Artifacts

FigJam board, live demo, and workflow examples.