Sr. Software Engineer, Generative Models

Intangible is a spatial intelligence company building a new class of creative tools.

Our flagship product, Intangible Studio, is a browser-based 3D platform that takes teams from idea to finished scenes and shots with speed and control. We combine real-time 3D with generative AI, so creatives keep hands-on control of composition, camera, animation, and the overall artistic vision. It is the one platform that brings LLMs, image, video, and 3D generation into a single workflow, backed by a multi-agent architecture and a proprietary world model. What you set up is what you get – the speed of AI visualization and the control of 3D.

Intangible runs in production with customers across film, advertising, retail, events, aerospace, automotive, and deeptech – high-stakes creative work where you can’t afford AI slop. Built by veterans of Apple, EA, ILM, Pixar, Unity, and BMW.

You will integrate new generative models (images, video, Gaussian splats) into Intangible Studio and contribute to features in our 3D engine.

New image, video, and 3D models ship constantly. Your job is to evaluate them quickly, find what is actually usable, understand new capabilities, and integrate the good ones into our pipeline within days of their launch, not months. That pipeline runs from a 3D scene to scene context to rich prompting and positional data, then into a model that returns a result in a specific style. You will work across the job system that powers it (TypeScript and Node on Google Cloud) and the visualizer that renders it, evolving both as model inputs get richer: image-to-image, image-to-video, video-to-video, multi-image references, and whatever lands next.

The bar is cinematic because our customers have high-stakes, real world use cases. We build for high-end creative work, so "it matches the scene composition" is table stakes. The harder question is whether the output actually looks good or reads as AI slop. You need the visual eye to tell the difference and the engineering range to fix it.

• Evaluate new image, video, and 3D models as they ship. Kick the tires, find what is usable, and integrate the winners end to end.

• Build and evolve the job system (TypeScript, Node, Google Cloud) as model inputs grow richer: image-to-image, image-to-video, video-to-video, and multi-image or video references.

• Own integrations all the way through: tuning hooks, backend wiring, and the visualizer and UX changes that ship with them, alongside other engineers.

• Pitch in on the 3D layer: scene composition, cameras and projection, WebGL and WebGPU, Gaussian splats, and shaders for post-process effects.

• Hold a cinematic bar on every output. Push back when something looks off instead of shipping it.

• Give the team reliable US-timezone coverage alongside 3D engineers based in Europe and Asia.

• Engineering judgment. You understand the code you ship and do not auto-accept AI-generated commits you cannot maintain.

• A real visual eye. You can explain why an image works or does not, not just that it looks fine. A photography, film, or animation habit is a good sign.

• Collaborative by default. You jump on a call or ask a question rather than spend three days proving you can do it alone.

• Excited by the craft. You see what these tools make possible and want every frame hitting the highest quality.

• Ahead of the curve. You keep us setting the pace instead of scrambling once something has slipped.

• 5+ years building and shipping production software, senior enough to own systems end to end

• Strong TypeScript and Node. TypeScript is around 90% of our codebase, so this one is non-negotiable.

• Backend platform experience with microservices on a major cloud. We use Google Cloud; AWS or Azure experience translates.

• Working knowledge of 3D: how cameras, projection, perspective, and depth of field actually work, plus WebGL or WebGPU

• Enough front-end ability to take a feature all the way to the UI

• A demonstrable eye for visual quality, ideally backed by a personal obsession with photography, film, animation, or 3D

Day one:

• TypeScript and Node.js (around 90% of our codebase)

• Microservices on a major cloud (we use Google Cloud; AWS or Azure translates)

• Three.js, WebGL and/or WebGPU

• 3D rendering fundamentals: cameras, projection, perspective, depth of field

• React.js to take a feature all the way to the UI

Nice to have:

• Gaussian splats

• Shaders for post-process work

• ECS (entity component system) patterns

• Fine-tuning or training generative models

• Generative pipelines (e.g. ComfyUI)

• Agentic AI or workflow tooling

• Microservices on a major cloud (we use Google Cloud; AWS or Azure translates)

• Game development experience