Engineering @ Lica

We are building a custom file format—one that is inherently agentic and design-focused. Most traditional systems were built for user clicks, not agents. We believe the future of design requires structured, composable formats that AI can reason about natively, while still being fluid enough for human creativity.

Think of us like the early OpenAI engineering team before ChatGPT launched: small, fast-moving, experimental, and obsessed with research-to-product velocity. But unlike then, our models are already in the hands of top brands across the U.S., running in human-in-the-loop workflows that power real marketing campaigns, presentations, and creative workflows every day.

Everyone Is a Full-Stack AI Engineer

At Lica, research, engineering, and product are strongly integrated. We don’t treat them as separate functions - we build them together, in the same loop:

• Challenges in graphic design rendering → You’ll tackle problems like building a renderer that keeps semantic structure intact (grids, typography hierarchies, brand constraints), ensuring round-trips between human edits and agent edits remain seamless, and making rendering pipelines fast enough to support real-time, multi-agent workflows.
• Deployment Infra → You’ll scale GPU-heavy training runs, optimize distributed systems, and build serving infrastructure for real-time interaction.
• Product & Agents → You’ll prototype and ship agentic workflows—turning raw model outputs into brand-aware ads, decks, videos, and campaigns.
• Customer Impact → You’ll sit close to the humans-in-the-loop—watching how real designers and marketers use the system, and then shaping features that remove friction and spark creativity.