The position is based at Microsoft Research Cambridge within the Future AI Infrastructure group. This role offers the opportunity to conduct cutting-edge research at the intersection of network protocols, systems design, and AI infrastructure, with direct relevance to real-world systems at global scale. You will…
By transforming remote sensing data into actionable insights for sustainable development, geospatial AI (GeoAI) is rapidly emerging as a vital tool to address the mounting challenges and urgent needs in Africa. These include climate change, rapid urbanization and food insecurity, which affected 307 million people…
What does it truly mean for a language model to “reason”? Current evaluations reward models’ correct standalone answers—but correctness alone reveals little about the process that produced them. We argue that reasoning should be understood not as a static chain of steps but as a…
Given that Large Language Models (LLMs) are increasingly applied to automate software development, comprehensive software assurance spans three distinct goals: regression prevention, reactive reproduction, and proactive discovery. Current evaluations systematically overlook the third goal. Specifically, they either treat existing code as ground truth (a compliance…
Indirect prompt injection attacks threaten AI agents that execute consequential actions, motivating deterministic system-level defenses. Such defenses can provably block unsafe actions by enforcing confidentiality and integrity policies, but currently appear costly: they reduce task completion rates and increase token usage compared to probabilistic defenses.…
OneDrive & SharePoint (ODSP) Applied Science’s mission is to invent the AI‑native knowledge substrate for Microsoft 365—turning the world’s largest enterprise content platform (ODSP) into a foundation for specialized, enterprise‑ready, planet‑scale AI and agents. We are seeking a passionate, creative, and analytical individual contributor with expertise…
We are looking for a Principal Applied Scientist who is creative, self-driven, curious, people-oriented, and comfortable defining a path through ambiguity toward high-level goals. Research areas of interest include but are certainly not limited to reinforcement learning, agentic systems multimodal modeling, post-training techniques and their application,…
We are seeking a Senior Applied Scientist with experience in working with LLMs who has designed scalable systems using LLMs. As a Senior Applied Scientist, you will design and build systems that allow LLMs to reason over large amounts of data as well as leveraging…
We are seeking a Senior Applied Scientist to join the Enterprise Voice Calling team within the Teams Calling, Meeting & Devices group. This role is ideal for someone passionate about building intelligent, scalable, and secure voice communication experiences powered by cutting-edge AI.
We are looking to hire a Senior Applied Scientist to join the science team responsible for some of the Artificial Intelligence innovations and experiences being delivered within Microsoft Teams and related products. The team works on state-of-the-art AI models and leverages innovation in their architecture,…
As a Principal Applied Scientist at Viva Engage, your role is to build Agentic applications across our product, predictive models that connect individuals with each other and knowledge across organizations. Our focus lies in discovery, recommendations, relevance and grounding for features across Viva Engage, leveraging…
As a Senior Machine Learning (ML) Engineer on the Viva Engage Search team, your experience in Search, Ranking, and Retrieval systems will help shape the future of intelligent search experiences in the Copilot and Agentic AI space. You will play a key role in designing,…
