Applied Sciences Manager , Ads Brand Safety and Suitability

Applied Sciences Manager , Ads Brand Safety and Suitability Overview

This role leads applied science work for Amazon Ads focused on protecting advertisers from unsafe, unsuitable, or policy-breaking content across web, mobile app, connected TV, and audio inventory. The team’s purpose is to make sure ads shown through Amazon’s demand-side platform appear next to content that aligns with advertiser expectations for trust and suitability, while also giving brands flexible controls to define what they consider acceptable. The work sits at the intersection of advertiser trust, publisher quality, and supply integrity.

The problem space is evolving quickly because generative AI has made it easier and cheaper to produce deceptive, low-quality, synthetic, and policy-evading content at scale. Traditional classification methods are no longer enough. This position is intended for an applied science leader who can guide the next generation of AI-based brand safety and content classification systems that operate at internet scale, make decisions in milliseconds, and continuously adapt to new forms of risk.

The team will be responsible for science and modeling efforts across several areas, including large language model-based classification and semantic understanding, real-time multimodal content evaluation, adversarial machine learning and model resilience, proactive risk intelligence and content risk discovery, detection of AI-generated and synthetic content, and identification and disruption of abusive content systems at scale. The broader goal is to define how modern AI can separate high-quality advertising inventory from unsafe or unsuitable content across all major ad surfaces.

What makes this opportunity distinctive

This is not a static classification problem. The environment changes constantly, and the systems must be able to reason about context, adapt quickly, and generalize beyond previously observed patterns of abuse. The role requires building living systems that learn from new threats in real time rather than relying on fixed models that quickly become outdated.

The work involves low-latency machine learning and LLM-powered systems that evaluate content safety, brand suitability, misinformation risk, and other emerging threats across massive live traffic streams. These systems must make billions of decisions every day while meeting single-digit millisecond latency requirements. The position combines frontier AI research, large-scale production engineering, and business-critical impact, because the models produced by the team directly affect major advertising spend and the trust of leading brands using Amazon DSP.

Core scientific challenges

The role involves difficult technical problems such as detecting sophisticated synthetic or AI-generated content, interpreting nuanced contextual brand risk, identifying coordinated MFA-space behavior before it scales, and designing systems that balance precision, recall, latency, explainability, and fairness. It also requires building adaptive models that can withstand adversarial changes and using LLMs for semantic understanding in environments where response time is tightly constrained.

More broadly, the position offers the chance to work where advanced AI, internet-scale systems, adversarial behavior, and meaningful business outcomes meet. The systems built by this team will help shape what trustworthy digital advertising infrastructure looks like over the coming years.

Responsibilities

• Define the vision, mission, and long-term strategy for applied science solutions that strengthen key parts of the contextual advertising product.
• Own the strategy and technical roadmap for machine learning and large language model-based classification systems.
• Monitor the contextual advertising landscape and identify algorithm investments that can keep Amazon’s solutions best in class.
• Lead a cross-functional organization of applied scientists and software development engineers, and grow a high-performing team focused on brand safety and AI-driven risk intelligence.
• Recruit, coach, and develop senior scientists while increasing the team’s speed of innovation and overall effectiveness.
• Build a team environment that values innovation, scientific rigor, rapid execution, and long-term thinking.
• Drive delivery from research and experimentation through production deployment at a scale of billions of classifications per day.
• Create scalable and automated processes for large-scale data analysis, model development, model validation, and model implementation.
• Use machine learning and statistical methods to design new solutions that can operate efficiently at very large scale.
• Advance multimodal understanding, semantic reasoning, and adaptive learning systems.
• Develop proactive detection and risk-hunting capabilities to surface new abuse trends and emerging threats early.
• Stay informed about new developments in machine learning and AI and convert promising ideas into production-ready solutions for Amazon Advertising.
• Shape organization-wide generative AI direction and communicate the team’s technical strategy to senior leadership.
• Partner with Product, Policy, Ads Quality, and Infrastructure teams to deploy AI capabilities in production at scale.
• Work closely with engineering and product management partners to advocate for new algorithms and implement complex machine learning models in live systems.
• Use business analytics and data science investigations to inform major business decisions and guide the algorithm roadmap.
• Represent the team’s innovation through peer-reviewed publications and whitepapers.

Requirements

• A master’s degree or higher in computer science, mathematics, statistics, machine learning, or a similar quantitative field, or an equivalent PhD-level background.
• Experience leading science teams.
• Experience building, launching, and operating AI products at scale.
• Strong understanding of machine learning and large language model fundamentals, including model architecture, training and inference lifecycles, and how to optimize model execution; alternatively, experience leading and influencing a team or organization in these areas.
• Experience working with both technical and business stakeholders in global, cross-functional settings.
• Experience leading large technical or engineering programs, with a track record that includes thought leadership, business case development, delivery of customer value, and successful completion.
• Preferred background in applied research.
• Ability to work on highly complex problems involving brand safety, suitability, misinformation, synthetic content, adversarial robustness, and low-latency multimodal classification.
• Comfort operating in a fast-moving environment where models must adapt to new risks and evasive tactics quickly.
• Experience balancing competing goals such as precision, recall, latency, explainability, and fairness in production systems.

Additional information

Amazon describes itself as an equal opportunity employer and says it values a diverse workforce. Hiring decisions are based on experience and skills. The company also states that it is committed to protecting candidate privacy and data security.

If you need a workplace accommodation or adjustment during the application, interview, or onboarding process because of a disability, Amazon directs applicants to its accommodations support resources for help.

The posting also notes that Amazon does not discriminate on the basis of protected veteran status, disability, or other legally protected status.

Applicants can use Amazon’s online application process to apply for the role. The page also links to Amazon’s privacy notice and accommodations information for candidates who need support during hiring.

Degree Requirement: Bachelor’s

Visa Sponsorship May be