Job Description:
Microsoft is a place where engineers and innovators collaborate to shape the future of technology and grow their careers beyond what is possible elsewhere. In a cloud-first, AI-driven world, we build platforms that are open, scalable, and designed to solve real business problems at global scale. The Business Central Development Engineering team, part of Microsoft’s broader Business & Industry Copilots organization, is responsible for building and evolving Microsoft Dynamics 365 Business Central—one of the world’s leading SaaS ERP solutions for small and mid-sized businesses. Our team sits at the core of how modern business applications are architected, engineered, and delivered, combining deep enterprise domain knowledge with cutting-edge cloud and AI technologies. We are looking for a Applied Scientist to join our team at a pivotal moment. This role offers a unique opportunity to help shape the future of AI-native ERP by leading the development of intelligent, agent-first capabilities embedded directly into Business Central. You will work on highly strategic initiatives that advance the vision of the autonomous enterprise, designing AI agents that reason across business data, workflows, and user intent to deliver measurable improvements in accuracy, performance, and cost efficiency. In this role, you will drive rapid experimentation and evaluation, develop robust metrics and frameworks for assessing agent behavior, and apply advanced techniques—such as reinforcement learning and multi-step reasoning—to improve decision-making in complex enterprise scenarios. You will collaborate closely with development engineers, product managers, and partner teams to ensure solutions are secure, reliable, performant, and extensible, enabling both Microsoft and the partner ecosystem to build confidently on the Business Central platform. The ideal candidate brings strong expertise in natural language processing and large language models, with hands-on experience in model evaluation, fine-tuning, and optimization. Experience with instruction tuning, reinforcement learning from human feedback (RLHF), prompt and context engineering, and tool-augmented generation is highly valued. You should be comfortable working in an experimentation-driven engineering environment, using both offline and online evaluation methods to iterate quickly and improve real-world outcomes. A solid understanding of enterprise application constraints—such as correctness, compliance, performance, and trust—is essential. Microsoft’s mission is to empower every person and every organization on the planet to achieve more. As part of this team, you will contribute to a culture rooted in growth mindset, collaboration, and accountability, while building inclusive, impactful technology that helps businesses thrive in an AI-first future.
