Introduction
Mirela Palescu pinned the first AI-generated pattern to a piece of cotton and saw immediately that the image was lying. On screen it looked rich and symmetrical. On fabric, the repeat broke, the colours muddied, and one motif sat exactly where a seam would cut through it.
Mirela lives in Vienna-Favoriten and now works between a sewing machine, a small digital portfolio and workshops for women who want to connect craft with AI.
Story of the Path into AI
For years Mirela had sewn clothes for neighbours. She knew how a fabric falls, how a pattern behaves on a body and why an idea that looks beautiful flat can fail in motion. When a customer mentioned a generative image tool, Mirela first used it for apron fabrics. It produced attractive images and almost no usable designs.
That annoyed her into learning. She had no design degree and was embarrassed by her German in public courses. At home she experimented with prompts that included material, repeat size, colour limits and printing constraints. Her daughter helped set up a small portfolio. The first series combined traditional-looking motifs with digital variants, but every design had to pass a real sewing test.
One early model invented a glossy texture that could not be printed on the chosen cloth. Mirela kept the failed sample. It taught her that a picture is not yet a textile.
Current Work
Today Mirela works with small labels and theatre groups. For one costume project, AI generated variations; Mirela decided according to drape, stage light and whether an actor could move in the garment. In her courses she starts not with prompts but with fabric samples on the table.
The work remains small, and she prefers that. A household side income has become an atelier where informal craft knowledge is treated as knowledge. AI speeds up variations, but the last test still happens under fingers, scissors and light.
Personal Advice
“Do not let anyone tell you handwork is old-fashioned. It is a good compass for AI,” Mirela says. She advises learners to touch the material before trusting the image.
Key Facts
Age and place: 47, Vienna-Favoriten.
Background: housework, migration experience, informal textile craft.
Entry into AI: digitally drafted fabric patterns checked through real sewing samples.
Focus today: AI and textile design.
Typical tools: generative image models, digital portfolio, fabric tests.
Werkstattnotiz
The red thread from Mirela’s first failed sample lies in a small tin. It reminds her that an image becomes a design only when it meets material. In the next course she plans to have participants deliberately create a beautiful but unprintable pattern.