Join Microsoft's AI Economy Team as a Senior Applied Scientist and contribute to building the core infrastructure that enables organizations to ground, deploy, and scale AI applications. This role focuses on developing sophisticated detection models to safeguard a large-scale, developer-facing platform from fraud and abuse. You will tackle challenges like identifying single actors behind multiple accounts, distinguishing legitimate users from fabricated ones during signup, and monitoring account behavior post-activation. Your work will involve navigating sparse and delayed labels, adapting to evolving abuse patterns, and ensuring rigorous evaluation before deployment.

Microsoft is committed to empowering every person and organization globally. We foster a culture of growth mindset, innovation, and collaboration. Our values of respect, integrity, and accountability drive an inclusive environment where everyone can thrive. Starting January 26, 2026, US-based employees within a 50-mile commute of a designated Microsoft office, and non-US employees within a 25-mile commute, will be expected to work from the office at least four days per week, subject to local laws and variations.

Develop and deploy identity-clustering models leveraging similarity, graph, linkage, and embedding techniques. Construct applicant-risk models using structured and unstructured data from identity, business, and behavioral signals. Build robust behavioral detection systems for active accounts through sequence modeling, anomaly detection, query-distribution analysis, and cross-account signal utilization. Train, fine-tune, and validate detection models, including feature engineering and adversarial testing against attacker-controlled inputs. Address challenges such as weak supervision, delayed and noisy labels, class imbalance, model calibration, operating-point selection, and drift. Investigate live abuse patterns, conduct offline and online experiments, and translate findings into reusable detection mechanisms and measurable coverage enhancements. Collaborate with product and engineering teams to productionize model decisions and provide critical technical analysis for privacy and legal reviews.

A Bachelor's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or a related field, coupled with 4+ years of relevant experience in areas like statistical predictive analytics or research, is required. Alternatively, a Master's Degree in a similar field with 3+ years of experience, or a Doctorate with 1+ year of experience, is acceptable. Equivalent experience will also be considered.

Preferred qualifications include 5+ years of experience in developing production-ready machine learning models, with proficiency in languages like Python, R, or Scala. Demonstrated expertise in adversarial machine learning, anomaly detection, or bot and automation detection is highly valued. Experience with API abuse, scraping, or model extraction as a defender, along with familiarity with privacy and compliance constraints on identity signals, is advantageous. Deep knowledge in areas such as graph neural networks, embedding-based entity linkage, sequence and time-series models of account behavior, or LLM-assisted enrichment is a plus. A solid understanding of experimental design, causal inference, and model calibration, including the selection and explanation of production operating points, is also preferred.

This role requires the ability to meet Microsoft, customer, and government security screening requirements, including a Microsoft Cloud Background Check upon hire and every two years thereafter.