Lica is building the first foundational models for graphic design—models that understand layout, typography, animation, and brand identity as core elements, not side effects. Our AI agents generate brand-compliant visuals, videos, and presentations with creative precision and structural coherence.

We're training large-scale multimodal models built specifically for the needs of design and marketing teams. Unlike generic vision-language models, ours are grounded in the nuances of brand expression and content layout. We’ve built our own data labeling pipeline and QA team, and are assembling what is rapidly becoming the largest curated design dataset in the world. This gives us a unique opportunity to build models that go far beyond templates—models that reason about structure, story, and visual impact.

As a researcher at Lica, you’ll work on developing novel model architectures, defining internal benchmarks, and pushing forward the state of multimodal design AI. You’ll have the opportunity to publish your work, contribute to open-source model releases, and lead innovation in an emerging domain. Your research won’t stay in papers—it will power real-world creative workflows, in production, at scale.

If you're excited to define a new category of foundational models and want to lead the AI movement in design, email on purvanshi@lica.world.