Zoia Shpak
Creative Director · AI Video Producer · Writer
WRINKLES Are Stories
In the 1980s, Spanish designer Adolfo Domínguez wrote "La arruga es bella" — the wrinkle is beautiful. He was talking about crumpled linen, and it was a small scandal. You were supposed to iron things.
So I took the sentence literally and made a series of black and white portraits.
Not "wrinkles are acceptable."
Not "age can be beautiful too."
Not "aging is normal."
The wrinkle is beautiful.
Wrinkles are Stories. Memories. Maps. Journeys. Proof.
Fill it, lift it, freeze it. Do whatever you want with your face.
Just leave something. One line the body wrote itself.
Wrinkles Are Stories — concept, art direction, AI generation, grading and typesetting.
Full series on Behance — link in the comments.
1
0
17
Water Payback Calculator
The Problem
Hotels, gyms, and apartment buildings overpay for water every month, and most owners never see it. Efficient showerheads are among the cheapest upgrades a building can make, but the upgrade sits undone: the only estimates available come from the companies selling the fixtures, and working it out yourself means an afternoon in a spreadsheet. So nobody does.
The Solution
The calculator answers the question in seconds. Four fields in — monthly water bill, number of showerheads, rate per cubic meter, property type. Two numbers out — annual savings, and how soon the upgrade pays for itself. The efficiency assumptions are my own, calibrated against EPA WaterSense showerhead standards, and they're shown in the app — a payback figure is only as good as the numbers behind it. The estimate runs off the bill; the tariff is collected for context and will drive the calculation once the model moves from bill-based to consumption-based.
Example: a hotel with 12 showerheads and an $800 monthly bill at $3/m³ saves $3,360 a year and pays back the upgrade in three months.
How It Was Built
I built it in Bubble in two days. I'm a writer and visual artist, not a developer — which meant I spent more time on where the numbers come from than on making them appear. Savings rates differ by property type: 45% for gyms, 35% for hotels, 25% for apartment buildings.
Design — one detailed prompt gave me a foundation I genuinely liked; the Agent is strong at structure and interface. From there I refined it by hand: fewer elements, more room, one moment of emphasis on the number that actually matters.
Data — two types: Calculation, which stores each set of inputs with its results, and EmailSubmission for the report requests.
Logic — this is where I took over completely. I wrote the formulas field by field myself in the visual editor. The Agent built the shell; the calculations are mine.
Bubble got me from idea to working app without writing a line of code. The visual formula editor is what made the calculations possible for me at all — I could see the logic instead of having to know the syntax.
What Works Today
End to end: input validation, live calculation across all three property types, the result view, and email capture with on-screen confirmation. The PDF delivery itself is the next release — the app collects requests, it doesn't send yet.
What's Next
Showerheads are the narrowest version of this question, and I kept it that way on purpose: a tool that answers one question well is more useful than one that asks for twenty numbers. The next version widens the scope.
2
116
I’m taking part in the CapCut AI Festival 2026 — Commercial category — with my fictional luxury jewelry campaign, “ZM — Worth the Wait.” 🎬
I created ZM specifically for the festival: the brand, the pendant, and the signature black leather wedge-shaped case that became part of the house’s visual identity.
The film is built around anticipation.
A woman waits in silence while a man rushes through the rain carrying the ZM case. Their parallel journeys converge at her door, where the case is finally opened and the pendant is revealed in color against an otherwise monochrome world.
The final image of a crowd waiting outside the ZM jewelry house reframes the story: she was not the only one waiting.
The wait itself becomes part of the value of the object.
🛠️ How it was made
The film was created primarily in CapCut Video Studio, which I used for AI video generation, editing, pacing, sound design, and final assembly.
CapCut’s AI Agent helped me develop the shot breakdown and translate my existing concept into production prompts and scene plans.
Gemini/Omni was used for a small number of additional shots.
The project includes the final commercial and a short Behind-the-Scenes look at the AI production process.
ZM — Worth the Wait.
0
32
First Date
Spec Ad
Role: End-to-End Production (Concept & Script, AI Generation, Sound Design, Editing).
Tools: Gemini Omni, CapCut Pro.
1
45