Overview

We’re looking for a Senior Applied Scientist with deep expertise in Applied AI, Machine Learning, and Deep Learning, and a proven track record of delivering innovative, user-facing AI experiences at scale. This is a unique opportunity to shape the future of enterprise productivity through cutting-edge Large Language Models (LLMs) and AI-powered experiences in Microsoft 365 Copilot, one of Microsoft's flagship products.

You will join a fast-paced, highly collaborative team of scientists, engineers, and product managers working at the forefront of AI innovation. In this role, you will drive the development of advanced LLM-based capabilities, reasoning systems, and real-time orchestration platforms, translating state-of-the-art research into impactful experiences for millions of enterprise users. Your work will directly influence product strategy, model quality, and the next generation of intelligent productivity solutions across Microsoft 365.

Microsoft’s mission is to empower every person and every organization on the planet to achieve more. As employees we come together with a growth mindset, innovate to empower others, and collaborate to realize our shared goals. Each day we build on our values of respect, integrity, and accountability to create a culture of inclusion where everyone can thrive at work and beyond.

Responsibilities

• Develop scientific approaches for evaluating and improving LLM reasoning, grounding, retrieval, and agentic behaviors using offline benchmarking, online experimentation, and large-scale telemetry analysis.
• Drive complex technical investigations, model evaluations, and experimentation to improve quality, reliability, latency, cost efficiency, and user experience.
• Define and execute rigorous evaluation methodologies, offline benchmarks, online experimentation frameworks, and data-driven approaches for measuring model and product performance.
• Partner closely with engineering and product teams to translate research innovations into scalable, production-ready systems used by millions of customers.
• Deeply analyze model behavior, failure modes, grounding effectiveness, reasoning quality, tool use, retrieval performance, and orchestration outcomes to identify opportunities for improvement.
• Develop and advance approaches for agentic workflows, multi-step reasoning, planning, retrieval-augmented generation (RAG), tool orchestration, and reinforcement-learning-based optimization.
• Review technical designs, experimental results, and implementation approaches to ensure high standards of scientific rigor, reproducibility, and engineering excellence.
• Collaborate with researchers, applied scientists, and engineers to solve challenging technical problems spanning LLMs, machine learning, deep learning, information retrieval, and human-AI interaction.
• Stay current with advances in Large Language Models, reasoning systems, agent architectures, reinforcement learning, and applied AI, incorporating relevant innovations into production systems.