Senior Applied Scientist

The AI Platform and Tools team in the Windows Platform and Developer organization builds end-to-end systems for AI inference and agent workflows on Windows. Our work makes Windows a versatile and powerful platform for on-device AI, enabling compelling experiences for customers and providing developers with the tools they need to build the next generation of intelligent applications. We also work on technologies that move work seamlessly between on-device and cloud-hosted models so experiences can deliver the right balance of quality, latency, privacy, reliability, and cost. We are looking for a Senior Applied Scientist to develop and ship machine learning innovations across the Windows AI stack. In this role, you will research, prototype, evaluate, and productionize techniques that improve model quality and the efficiency of AI workloads on a diverse range of Windows devices. You will work at the intersection of applied machine learning and systems, partnering with software engineers, hardware architects, program managers, and researchers to solve challenges in model optimization, search and retrieval, inference orchestration, and agent execution. This is an opportunity to turn advances in generative AI, ML/algorithms for search and retrieval, and agentic systems into platform capabilities used by developers and customers at Windows scale. 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.

Bringing the State of the Art to Products Partners with Engineering and Product teams to turn advances in generative AI, search and retrieval, agentic systems, and efficient inference into measurable product impact. Builds prototypes and production-ready platform components for inference, retrieval, and agent workflows across heterogeneous CPUs, GPUs, and NPUs on Windows devices. Leveraging Applied Research Develops and evaluates data-, research-, and experimentation-backed techniques for on-device and hybrid inference, including quantization, distillation, model adaptation, compression, indexing, embeddings, and hardware-aware optimization. Investigates intelligent model and workload placement across device and cloud resources while balancing quality, latency, memory, power, reliability, privacy, and cost. Machine Learning Functionality, Insights, and Technical Tools Designs datasets, metrics, experiments, and benchmarks for model and system evaluation; analyzes behavior to identify quality and performance bottlenecks and drives improvements across the AI workload lifecycle. Implements and integrates machine learning components, runtimes, developer APIs, and tools, then validates their behavior through production-oriented testing and monitoring on representative hardware and workloads. Documentation Documents scientific approaches, experiment plans, datasets, evaluation results, design decisions, and implementation guidance so that work can be reproduced, reviewed, and adopted by partner teams. Communicates findings through design reviews, technical presentations, and, where appropriate, patents or publications. Ethics and Privacy Applies Responsible AI and Microsoft security principles when selecting data, designing experiments, and developing on-device, hybrid, retrieval, and agent systems. Identifies risks involving privacy, security, bias, and reliability and incorporates appropriate safeguards into technical solutions. Capability Management and Networking Mentors engineers and product partners on applied machine learning, evaluation, inference optimization, search and retrieval, and agent workflows. Builds collaborative relationships across science, engineering, hardware, and product