Kat Kampf

Kat Kampf is a Google DeepMind product manager working on vibe coding in Google AI Studio, where she helps people turn natural-language instructions into functioning Gemini-powered applications. Her career spans Microsoft, GitHub, an AI marketing startup, and Google, with a consistent focus on making sophisticated technical tools easier to use.

From developer tools to AI-powered creation

At Microsoft, Kampf worked on Data Accelerator for Apache Spark, an open-source system designed to simplify enterprise data processing; she detailed its 2019 improvements as a program manager. At GitHub, she helped integrate GitHub Classroom with Visual Studio Code, bringing assignment setup, editing, version control, grading, and collaboration into a more cohesive experience for students and educators.

She subsequently worked on Blaze.ai, an AI marketing product for entrepreneurs and creators, before joining Google's AI and machine-learning developer team. In late 2024, she published open-source Gemini starter apps for spatial understanding, video analysis, and mapping. Her subsequent Google AI Studio work introduced native code generation and agentic tools and a redesigned application-building experience.

What she is building toward

• Prompt-to-app development: Describe the intended outcome, and the development environment assembles an initial application, connects relevant Gemini capabilities, and supports refinement through editable examples and visual iteration.

• Context-aware multimodal applications: Gemini can combine image generation, live interaction, and Google Search or Maps grounding. Kampf has demonstrated search-informed illustrations and a webcam-based tennis coach that gives immediate feedback on a player's swing.

• A path beyond disposable prototypes: Browser-generated applications should have a route into more capable development environments. Kampf has prototyped exporting AI Studio projects to Google Antigravity and described newer capabilities including multiplayer applications, databases, authentication, and external APIs.

At AI Engineer Code 2025, she articulated the larger ambition: intent-driven full-stack development, in which an application builder can infer requirements such as storage or payments without expecting newcomers to specify every underlying infrastructure component.