A public lab for systems that deserve a closer look.

I'm Nini. I build software, study how AI-enabled systems fail, and document the parts that are useful beyond the experiment.

DVNLL Labs connects my AI-security learning, local-model infrastructure and the process of building a continuous security service. Study notes are labelled as study notes; practical results include the method and evidence.

The authorship is personal. The professional security work sits under DVNLL. Trogawa remains its own product, appearing here only when a technical or learning-design decision is useful beyond that project.