Join Microsoft's AI Economy Team as a Principal Applied Scientist and play a crucial role in building the infrastructure that powers AI applications. This role focuses on Trust & Safety, developing sophisticated detection systems to combat abuse in AI access. You will be at the forefront of innovation, safeguarding AI platforms by verifying user identities, detecting coordinated malicious activity, and distinguishing legitimate usage from exploitation.

Microsoft is dedicated to empowering every individual and organization globally. Our culture thrives on a growth mindset, innovation, and collaboration, underpinned by values of respect, integrity, and accountability. We foster an inclusive environment where everyone can achieve their full potential.

Please note: Beginning January 26, 2026, employees within a designated commute distance of a Microsoft office in the U.S. or internationally are expected to work from the office at least four days per week. This guideline is subject to local laws and may vary by region.

Lead end-to-end AI Economy Trust & Safety projects, from problem definition and planning to production rollout and measurement. Take ownership of identity resolution and clustering strategies to link accounts, tenants, and payment instruments to common actors. Develop and manage applicant-risk and ongoing account-risk models, including the foundational evidence for initial decisions and subsequent re-evaluations.

Innovate detection methods to differentiate high-volume legitimate use from coordinated extraction and other abusive behaviors. Select and critically evaluate machine learning approaches, including supervised learning, anomaly detection, graph methods, sequence models, and large language models. Establish a robust evaluation framework for model quality, calibration, false-positive impact, drift, explainability, and adversarial robustness, while maintaining the associated threat model.

Collaborate closely with product and engineering teams within the AI Economy team to translate detection insights into actionable product controls. Coordinate effectively with privacy, legal, and external dependency stakeholders to ensure comprehensive risk mitigation.

A Bachelor's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or a related field, combined with at least 6 years of relevant experience in areas like statistics, predictive analytics, or research. Alternatively, a Master's Degree in a similar field with 4+ years of experience, or a Doctorate with 3+ years of experience is also suitable. Equivalent practical experience will be considered.

Preferred qualifications include proven experience in taking ambiguous, adversarial problems from initial framing through to deployed systems, including rigorous evaluation. Prior experience in know-your-customer (KYC), identity proofing, or payment risk within a self-serve or developer-facing product is highly valued. A track record of mentoring scientists and engineers, and elevating the technical standard of a team is also desirable.

Demonstrated depth across multiple modeling families (graph/network methods, anomaly detection, sequence models, LLMs), with clear evidence of their efficacy against adversarial data. Experience in reasoning about attacker economics and the cost-benefit analysis of controls is beneficial. Contributions through patents, peer-reviewed publications, or open-source projects in machine learning, security, or risk modeling are a plus. A minimum of 7 years of experience building and delivering fraud, abuse, security, or risk detection systems into production environments is required.