obliquus is an independent applied research lab building tools, systems, and models for problems that don't respond to the obvious approach.
most software and ai tools are built for the problems that are easy to see. we are interested in the ones that aren't.
the signal we care about lives in unstructured places: video, live audience behavior, artifacts that don't announce themselves. the straight line misses them by design.
obliquus builds research tooling and bespoke systems for teams working where the obvious approach stops being enough.
research on extracting structured information from video at scale. we focus on events, objects and actions that carry commercial or operational meaning — even when they appear incidentally, in unconstrained footage.
work on real-time understanding of live video together with its surrounding signal — comments, reactions, chat — as a combined source of insight. we study how meaning emerges from the interaction between broadcast and audience.
beyond ai, we take on systemic challenges that demand custom logic. if a problem involves deep tech, intricate data, or unconventional mechanics and defies off-the-shelf tools, we solve it.
we start with the data that's actually there, not the data that would be convenient. most oblique problems are defined by what their signal refuses to look like.
// change rarely · apply everywhere
coming soon
// a lab · not a startup
obliquus is an independent applied research lab. we build tooling, systems, and models for problems that resist the obvious approach — the ones that live off-axis, in unstructured signal, and in the seams between conventional frameworks.
for research collaborations, commercial inquiries, or to say hi —