Craft,engineered.
Senior Design Engineer · AI-Native
Interactive Experiences for AI-native Brands.
Building from chaos, where craft is a discipline.
Where I started
At NJIT I took a work-study job in a small room with the Instructional Technology Media Services Department. The role was helping put professors' course material online. That job led me to volunteer on a revamp of the department site, where I learned web, video, design, music, and photography.
That turned into adtech and fintech, building Fortune 500 campaigns at Undertone and the A/B frameworks behind them.
Now I build creative systems with AI, and the harness around the model is where the craft goes.
AI is only as good as the person steering it.
What I care about
Everybody has access to all the same tools, but nobody has access to the same way of thinking.
Explore. Generate.
Refine. Scale.
Working together, the AI produces the options and we decide which ones ship. Most of the time goes into exploring, and the rest splits between building and refining.
We can go further than we used to, and faster, with interactions and shader code as well as static layouts.
It rarely lands first try, so we curate and keep tuning the prompt and the inputs until it does.
We keep what's working, fix what's broken, and finish it.
Once something works, it becomes a system we can reuse.
Universal Context Protocol.
What I’m building now
Everything your brand sounds like, looks like, and stands for, in one place an AI can build from.
The one place that holds the voice, the design, and the rules we build everything from.
A job the system has been taught to do your way, ready whenever you ask for it.
The finished work, drawn from Brand Truth and on-brand the moment it arrives.
A validator checks every output against Brand Truth, and what it finds goes back into the source.
Brand systems and creative workflows, localized across markets.
Skincare productSkincare productSkincare productTaking each spot from idea to finished creative. The three above are one product ad, generated once and localized into English, Japanese, and Korean.
“You can outsource your thinking, but you cannot outsource your understanding.”
— kache, cited by Andrej Karpathy
Why I Care About Building the Trust Layer
Why I build AI-generated content, label it, and care about attribution, from adtech and fintech to First Day.
Read the articleProgress Through Trust, Creativity, and Collaboration
At Zeta Global, my team gave me room to figure things out. Together we took campaign output from one every two weeks to 3-5 a week, and I gave the people I trained the same room.
Read the article
Moving up the levels of AI, together.
How we work with AI
This half is how I work on a team. We use AI every day, and what one of us learns gets shared, so the whole team moves up a level together.
AI adoption has levels, numbered the way self-driving cars are. The comparison covers the numbering only. Each level is how much of a job runs without a person prompting each step. At level 1 someone pastes prompts into a chatbot. At level 5 a system runs whole jobs on its own. We move the work up one level at a time, so each workflow keeps running after the demo. The list shows levels 1, 3, and 5.
- L1Chats People prompt by hand, and each answer stays in the chat it came from.
- L3Workflows A task runs the same way every time, and a person reviews the result.
- L5Agents Whole jobs run end to end, and people keep the decisions.
Get in touch
Working end to end across design, creative, and engineering, where creativity, collaboration, and ownership build toward the company's goals.
jeffkliu@gmail.com