Back to all jobs
A
Senior Program Manager, Intellectual Property Protection
Vancouver, British Columbia, CAN
full-timeProject/Program/Product Management--Non-TechJob Description
Are you passionate about leveraging machine learning to protect Amazon's marketplace at scale? We're seeking a Senior Program Manager who will drive our brand protection detection systems through systematic problem-solving and technical expertise. In this role, you'll own the end-to-end program lifecycle for ML-powered detection systems, going 2-3 levels deeper than surface analysis to identify root causes and build solutions that eliminate entire classes of false positives. You'll work directly with applied scientists and engineering teams to translate complex abuse patterns into engineering-ready specifications that scale to hundreds of millions of product listings globally.
This position sits at the heart of Amazon's brand protection systems, where your decisions directly determine whether hundreds of millions of product listings are correctly identified as legitimate or infringing. You'll partner with applied scientists on ML model performance, manage golden dataset refresh cycles, optimize image enrollment processes, and conduct root cause analysis on false positive drivers. Your strategic impact will be felt across marketplaces, regions, and languages as you balance precision and recall at unprecedented scale.
Key job responsibilities
- Partner with applied scientists to evaluate machine learning model performance for brand protection detection systems, making high-judgment decisions that directly impact detection accuracy across billions of listings WW.
- Manage high-quality training dataset refresh cycles and text workstream pattern identification across multiple marketplaces and languages to ensure detection models have high-quality training data.
- Conduct comprehensive root cause analysis on false positive drivers, going 2-3 levels deeper than surface analysis to identify systematic improvements for detection systems.
- Translate findings into detailed engineering-ready specifications with clear, measurable success criteria that enable engineering teams to build to specification.
- Present strategic recommendations to Directors and VPs on detection system enhancements, establishing metrics and driving alignment across stakeholders.
A day in the life
Your day begins by reviewing overnight detection metrics across marketplaces to identify any emerging patterns requiring attention. You'll meet with applied scientists to evaluate a new model iteration, making judgment calls on its precision/recall balance before implementation. After, you might spend time analyzing text patterns across multiple languages to identify new detection opportunities or coordinate with engineering on implementing your latest false positive reduction blueprint. You'll wrap up by presenting findings from your latest root cause analysis to leadership, highlighting systematic improvements that could eliminate thousands of false positives while maintaining strong brand protection.
About the team
The brand protection program team serves as Amazon's systemic problem-solving engine, providing vertical product-program support across discovery, detection, and remediation systems.
Product and engineering teams explicitly prefer working with our team because we deliver detailed blueprints instead of vague problems, enabling engineering to build to specification and validate against clear success criteria.
This position sits at the heart of Amazon's brand protection systems, where your decisions directly determine whether hundreds of millions of product listings are correctly identified as legitimate or infringing. You'll partner with applied scientists on ML model performance, manage golden dataset refresh cycles, optimize image enrollment processes, and conduct root cause analysis on false positive drivers. Your strategic impact will be felt across marketplaces, regions, and languages as you balance precision and recall at unprecedented scale.
Key job responsibilities
- Partner with applied scientists to evaluate machine learning model performance for brand protection detection systems, making high-judgment decisions that directly impact detection accuracy across billions of listings WW.
- Manage high-quality training dataset refresh cycles and text workstream pattern identification across multiple marketplaces and languages to ensure detection models have high-quality training data.
- Conduct comprehensive root cause analysis on false positive drivers, going 2-3 levels deeper than surface analysis to identify systematic improvements for detection systems.
- Translate findings into detailed engineering-ready specifications with clear, measurable success criteria that enable engineering teams to build to specification.
- Present strategic recommendations to Directors and VPs on detection system enhancements, establishing metrics and driving alignment across stakeholders.
A day in the life
Your day begins by reviewing overnight detection metrics across marketplaces to identify any emerging patterns requiring attention. You'll meet with applied scientists to evaluate a new model iteration, making judgment calls on its precision/recall balance before implementation. After, you might spend time analyzing text patterns across multiple languages to identify new detection opportunities or coordinate with engineering on implementing your latest false positive reduction blueprint. You'll wrap up by presenting findings from your latest root cause analysis to leadership, highlighting systematic improvements that could eliminate thousands of false positives while maintaining strong brand protection.
About the team
The brand protection program team serves as Amazon's systemic problem-solving engine, providing vertical product-program support across discovery, detection, and remediation systems.
Product and engineering teams explicitly prefer working with our team because we deliver detailed blueprints instead of vague problems, enabling engineering to build to specification and validate against clear success criteria.