The Shift For the last decade, frontend engineering has largely been defined by deterministic patterns: click a button, fetch data, render a list. However, the era of static interfaces is coming to an end. At Luma AI, we are building multimodal AGI, systems that can see, hear, and imagine. This new form of intelligence requires a new syntax of interaction.

The Mission We are a full-stack AI lab with the capital and compute of a tech giant, yet we operate with the speed and leanness of a small startup. We train foundational models and build the products that deliver them. We are looking for a Design Engineer to invent the UI primitives for this new paradigm, bridging the gap between the unpredictability of generative models and the user’s need for control and creativity.

The Challenge

• Inventing Syntax: Define how humans collaborate with agents. How does a user interrupt a video generation stream? How do we visualize the "thought process" of a multimodal model? You will answer these questions with code.
• Engineering Trust: You will build interfaces that make complex, asynchronous AI workflows feel transparent and responsive, managing the delicate balance between latency and magic.
• Orchestrating Chaos: You will architect frontend systems capable of handling heavy media assets and non-deterministic states without compromising the visual fluidity and responsiveness of the application.

The Builder We Need

• A "Design Engineer": You don't just implement specs; you obsess over the physics of a toggle and the timing of a transition. You believe that in an AI-native product, the interface is the product.
• Master of the DOM: You have deep, production-level expertise in React and TypeScript. You understand the browser rendering pipeline intimately and know how to extract 60fps performance even under heavy computational load.
• Inventor: You are tired of building standard SaaS dashboards. You want to build 0-to-1 interfaces for a technology that is still being defined.