I came into AI through design, philosophy, and visual culture and my background still shapes how I build.

My work sits between applied AI research, product design, and evaluation. I’m building Polinko, an eval-led research model for reliable, ethical human-AI interaction, alongside Polinko Toy Factory, a smaller experimental lab for testing fun ideas, behavioural consistency, and method in practice.

I’m especially interested in signals: how products, prompts, runtimes, and evals can make model behaviour easier to observe, judge, and improve. The kind of work that holds up is conceptually weird and operationally serious: clear methods, clear evidence, and enough product shape to make the research real.

Your sponsorship will help keep my work independent and alive. It gives me room to build carefully, run experiments properly and keep sharing the systems, notes, and research that come from of it.