Software Engineer - Microsoft Research

Join a team working at the center of frontier AI.You'llcollaborate directly with researchers in one of the world's premier research labs, transforming breakthrough ideas into real, reliable, scalable capabilities used across Microsoft products. This role sits at the highimpact intersection of researchengineeringproduct, where the work you do directly shapes what AI becomes next.

Focus Areas You May Work On

• Agentic Memory

• Efficiency

• Security

• Frontier modelarchitectures

• Mathematics and logic applied to LLMs

You might be a great fit if you:

• Motivated by the intersection whereresearch meets engineering - the space wherenew ideasbecome real systems.

• Engaged effectively with ambiguous, open-ended problems where exploration and clarity are equally important.

• Motivated byperformance, safety, correctness, and experience, not just cleverness.

• Communicate clearly, collaborate naturally, and debuganything withcalm and rigor.

• Inspired to build systems that meaningfully advance what AI can do, not just iterate on what already exists.

Weoperatewith urgency and ambition:

• Short iteration cycles: hypothesis prototype evaluate refine ship

• Deep collaborationwith researchers inventing new capabilities

• High ownershipof systems, experiments, and outcomes

• Evidence-driven decisionsinformed by evals, metrics, and safety signals

• A builder's mindset: clarity, craft, and care for the details that make systems robust

You'lljoin a teammotivatedby curiosity, accountability, and the conviction that we are building something thattruly movesthe field forward, tackling problems thatdon'teven have names yet alongside people defining the state of the art. If this motivates you, we want to meet you. Bring your curiosity. Bring yourcraft. Bring your appetite for building the future.

Microsoft Research is committed to cultivating a workplace that reflects the diversity of the global communities we serve. We actively seek to build an inclusive and equitable research environment where individuals of all backgrounds feel valued, supported, and empowered to contribute. We encourage applications from candidates who bring a broad range of experiences and perspectives. Our team works collectively to ensure that Microsoft Research remains a respectful, collaborative, and high-impact space for all researchers.

Microsoft's mission is to empower every person and every organization on the planet to achieve more, and we are dedicated to this mission across every aspect of our company. Our culture is centered on embracing a growth mindset and encouraging teams and leaders to bring their best each day. Join us and help shape the future of the world

• Translate research into systems:take early concepts and turn them into robust prototypes and product-ready components.

• Design highvelocity experimentation loops:datasets, evals, benchmarks, agent flows, safety tests, and decision frameworks.

• Scale andoptimize:training, inference, efficiency, latency, cost, reliability.

• Partner crossfunctionally:coinventwith researchers,codesignwith product teams, and land features that matter.

• As a Research Software Development Engineer inMicrosoft Research, you will design, develop, execute, andimplement technology research projects in collaboration with other team members, research partners, and product groups.

• Work with the team toidentify, evaluate, and deploy new AI tools and technologies to accelerate our progress.

• Embody ourcultureandvalues.

Required Qualifications

• Bachelor's Degree in Computer Science, or related technical discipline with proven experience coding in languages including, but not limited to, C, C++, C#, Java, JavaScript, or PythonOR equivalent experience.

Preferred Qualifications

• Ability tocompetentlyoperateacross an entire modern AI-centric software stack.

• Adaptability to dive into a new domain quickly.Polyglot technologycuriosity.Be a voracious learner of new languages, paradigms, tools, and techniques, consuming publications.

• 1+ year(s) experience buildingagenticcodingsystems thatoperateatcloud scale using common agent protocols and coding frameworks.

• Systematic use of benchmarking and other data-driven techniques to guide design and development.

• Ability to thrive in the face of ambiguity, able to prototype starting with a research concept or paper rather than a spec.Bea thought partner andoperateas part of a multidisciplinary team as we build prototypes quickly, fail fast, pivot, and adapt.

• Experience with data science, machine learning (ML), Large Language Models (LLMs), Natural Language Processing (NLP), computer vision, multimodal input and processing, and softwaresecurity.