Trailr AI

An early-stage AI video platform that secured enterprise trials with Nordisk Film and DR. I joined as a frontend intern and, after graduating, continued part-time with equity warrants, owning the product side of a full UI/UX redesign.

The platform had grown feature-first and the UX had drifted. The real constraint: improve the product meaningfully without forcing backend changes a small team couldn't afford. The job was deciding what was worth building against what the existing backend could actually support.

Product judgment over raw output. Competitor research to find the gaps, ruthless scoping to ship within real constraints.

Benchmarked direct and indirect competitors, synthesized the findings into a single design direction, and scoped ruthlessly to the backend's limits, cutting features rather than forcing rewrites. When the navigation proved convoluted, I led a click-reduction overhaul, pulling Settings and Feedback out of the primary workspace so screening and building stayed front and centre. After the Nordisk Film pitch surfaced a need to make the AI less of a black box, I pushed for a thinking UI that streams the model's reasoning word by word, trading a flashy result for the transparency enterprise clients actually trust. Throughout, I worked directly with the founder and engineers to sequence what shipped.