UI Studio is Spotify’s internal generative UI prototyping platform that enables teams to spin up fully functioning replicas of Spotify surfaces — complete with playable music — to explore and pitch new product ideas. What started as a small internal project has evolved into a widely adopted platform, and we’re now expanding it into a full generative UI system where teams can prompt for the interfaces they need.

You’ll join a small, high-impact team working at the intersection of machine learning, design systems, and product development — helping shape how user interfaces are imagined and built at Spotify.

What You'll Do

• Shape the machine learning direction for UI Studio as it evolves into a generative UI platform
• Build systems that translate natural language into fully functional UI experiences through applied agent chat
• Design and implement agent-based architectures (e.g., Claude Agent SDK or similar) as the core interaction layer for UI generation
• Lead work on computational understanding of UI — transforming design files, product specs, and prompts into structured, machine-readable representations
• Develop pipelines that convert unstructured inputs (e.g., Figma files, free-form prompts, internal data) into structured UI blueprints
• Apply ML techniques to generate, interpret, and manipulate visual interface structures at scale
• Partner closely with design and product to bridge the gap between human intent and generated UI output
• Help integrate an evaluation framework into the application to help guide our system to produce better results
• Contribute to evolving the ML and agent tech stack across TypeScript and Python environments
• Help scale the platform from early adoption into a foundational capability used across Spotify
  

Who You Are

• You have strong experience in machine learning, with a focus on applied AI systems in production environments
• You bring deep experience in Computational Design — especially how structured representations can be derived from unstructured inputs
• You have worked on problems involving transforming text, design artifacts, or multimodal inputs into structured outputs
• You are experienced with agent-based systems and LLM orchestration frameworks (e.g., Claude Agent SDK or similar)
• You understand how to design applied agent chat systems, where natural language is the primary interface
• You are comfortable working with natural language parsing, intent extraction, and building systems that interpret human input reliably
• You have experience working with design systems, UI data, or adjacent domains (e.g., Figma APIs, front-end structures, or visual representations)
• You are experienced with modern ML stacks and can work across Typescript and/or Python
• You care about building intuitive, high-quality user experiences powered by ML, not just models in isolation
• You enjoy working in small, high-impact teams where ambiguity is part of the problem space
  

Where You'll Be

• This role is based in Stockholm
• We offer you the flexibility to work where you work best! There will be some in person meetings, but still allows for flexibility to work from home.