Applied Scientist, Ads Brand Safety and Suitability

Overview

In this role you will lead the science strategy for AI-powered brand safety at scale. You will build and operate LLM-powered classification systems that make billions of decisions per day with millisecond latency. You’ll enable real-time, multimodal evaluation across text, images, audio, and video to protect advertisers and elevate supply quality. You’ll bridge frontier AI research with production systems, shaping resilient models against adversarial content while driving measurable trust and business impact.

Responsibilities

• Own the science strategy for AI-powered brand safety classification
• Build LLM-powered content classification systems making billions of decisions/day at single-digit millisecond latency
• Develop multimodal evaluation pipelines reasoning across text, images, audio, and video in real time
• Design adaptive ML systems resilient to adversarial evolution, semantic understanding for nuanced contextual brand risk
• Define measurement frameworks and drive continuous improvement
• Translate research into production — own the path from prototype to deployed model
• Publish at peer-reviewed venues; contribute to the scientific community in adversarial ML, NLP, and content safety
• Collaborate with software engineering teams to integrate successful experiments into large-scale, highly complex Amazon production systems

Key requirements

• PhD, or a Master’s degree and experience in CS, CE, ML or related field
• Experience in patents or publications at top-tier peer-reviewed conferences or journals
• Experience programming in Java, C++, Python or related language
• Experience in any of the following areas: algorithms and data structures, parsing, numerical optimization, data mining, parallel and distributed computing, high-performance computing
• Experience in building machine learning models for business application

• LLM-powered content classification
• real-time, low-latency ML systems
• multimodal evaluation (text, images, audio, video)
• adversarial ML and model resilience
• semantic understanding of brand risk
• productionizing research into deployed models

…