Principal Applied Science Manager - Foundation Models, Agents & Trust Systems

Principal Applied Science Manager - Foundation Models, Agents & Trust Systems

Overview

About the Role

Microsoft Advertising serves ads across search, native, display, video, commerce, and emerging AI-powered experiences. Protecting these ecosystems requires decision systems that can understand advertisers, content, behavior, intent, landing pages, identities, and marketplace activity across every modality.

We are seeking a Principal Applied Science Manager to lead the science organization responsible for risk, editorial quality, moderation, policy enforcement, and Responsible AI across Microsoft Advertising.

The team will build the next generation of intelligent decision systems, including:

• Foundation models for advertiser behavior and risk.
• Foundation moderation models spanning text, image, video and multimodal experiences.
• Deep-research agents that investigate complex cases, gather evidence, reason across signals, and assist automated and human decision-making.
• Tiered enforcement systems that select the right model, workflow, and level of review based on risk, severity, confidence, latency, and cost.
• Human-in-the-loop that combine advanced AI capabilities with expert judgment and accountability.
• Continuous evaluation and audit systems that measure quality, fairness, safety, robustness, and business impact.

This role owns both scientific direction and production impact of the science. The successful candidate will define the long-term strategy, build and lead a strong science team, and work closely with engineering, product, platform, policy, review operations, and partner organizations to ship these capabilities across all Microsoft Ads products.

The role is ideal for a leader who can operate across multiple horizons: delivering measurable improvements today while shaping the future of Responsible AI, trust and safety, and intelligent enforcement systems.

Responsibilities

Responsibilities

• Define and drive the multi-year science strategy for risk, editorial quality, moderation, policy enforcement, and Responsible AI across Microsoft Advertising.
• Build, lead, and grow a high-performing team of applied scientists working across foundation models, multimodal understanding, behavior modeling, agentic systems etc
• Development foundation behavior models that understand advertisers, accounts, domains, identities, payments, content, and activity over time.
• Develop foundation moderation models that generalize across policies, products, languages, markets, and modalities.
• Build Deep Research agents that can investigate complex cases, retrieve and assess evidence, reason across multiple signals, identify contradictions, and support high-quality decisions.
• Design tiered enforcement architectures that combine lightweight classifiers, specialized models, foundation models, agents, deterministic systems, and human review.
• Determine when decisions should be automated, escalated to advanced models or agents, or routed to expert human reviewers.
• Establish scientific foundations for risk scoring, severity estimation, uncertainty, calibration, explainability, and cost-sensitive decision-making.
• Drive measurable improvements in user safety, marketplace integrity, advertiser experience, decision quality, operational efficiency, and revenue protection.
• Evolve scientific and engineering approaches as Responsible AI expectations, adversarial behaviors, policies, and model capabilities change.
• Work across product management, platform engineering, review operations, policy, legal, privacy, Responsible AI, and partner science organizations
• Influence senior leaders on scientific strategy, platform architecture, organizational investments, technical priorities
• Mentor senior scientists and managers, raise scientific standards, and build the next generation of applied-science leadership.

Qualifications

Required Qualifications

• Bachelor’s degree in Computer Science, Statistics, Electrical Engineering, Computer Engineering, or a related field and 15+ years of relevant experience;
  or a Master’s degree and 12+ years of relevant experience;
  or a Doctorate and 10+ years of relevant experience;
  or equivalent experience.
• Demonstrated experience leading applied-science or machine-learning teams and developing senior technical talent.
• Proven track record of defining scientific and product strategy and translating it into large-scale production capabilities with measurable customer, business, and operational impact.
• Ability to make complex product and technical trade-offs across quality, coverage, latency, cost, explainability, safety, and speed of delivery.
• Deep expertise in one or more of the following:
• Foundation models and large-scale representation learning.
• Fraud, abuse, risk, trust and safety, or cybersecurity.
• Content moderation, editorial quality, or policy enforcement.
• Multimodal understanding across text, image, video, audio, and web content.
• Agentic systems, retrieval, reasoning, and evidence-based decision systems.
• Large-scale classification, ranking, recommendation, or decision systems.

 

• Experience leading complex initiatives across engineering, product, operations, policy, and partner science organizations.
• Ability to lead the productionization of complex machine-learning systems, including data and labeling strategy, experimentation, model evaluation, deployment architecture, observability, reliability, latency, capacity, cost, and operational readiness.
• Ability to connect scientific advances with product requirements, operational workflows, engineering constraints, and business outcomes.
• Demonstrated ability to operate effectively in ambiguous and rapidly changing technical, regulatory, and Responsible AI environments.
• Strong communication and executive-influence skills.

Preferred Qualifications

• Experience building foundation models for behavior understanding, moderation, risk, or trust and safety.
• Experience with agentic systems, tool-using agents, deep-research workflows, retrieval, structured reasoning, and evidence-based decision systems.
• Experience designing multi-stage or tiered model architectures that balance accuracy, latency, coverage, and cost.
• Background in advertising, search, commerce, recommendations, financial risk, cybersecurity, or another high-scale marketplace domain.

This position will be open for a minimum of 5 days, with applications accepted on an ongoing basis until the position is filled.

Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations and ordinances. If you need assistance with religious accommodations and/or a reasonable accommodation due to a disability during the application process, read more about requesting accommodations.

Experience Level

Senior Level