You will set the end-to-end scientific direction for Excel's agentic AI and evaluation systems, translating customer needs, real-world usage patterns, and product failure modes into clear research, engineering, and product priorities - from early prototyping and experimentation through production deployment, evaluation, and ongoing service operation. You will partner closely with engineering, product management, customers, and Microsoft's broader AI, research, and platform communities to create trustworthy, high-quality AI applications that solve meaningful customer problems on a global scale.