Applied Science Manager, Seller Partner Trust and Store Integrity Science - Global Risk Intelligence and Prevention

Written by the T&S Network from Amazon's posting.

Leads a team of applied scientists building AI-powered seller risk detection for Amazon's marketplaces, catching bad sellers from registration through every marketplace action across 24 global marketplaces. The team sits in GRIP Science, the core detection engine within TSI Science, and works to stay ahead of adversaries who keep changing tactics.

What you'd do

• Set the science direction and multi-year plan for the seller detection area
• Decide which detection work to build now, which to defer, and how success is measured
• Weigh urgent requests from enforcement, policy, operations, and engineering against lasting fixes
• Write planning narratives for org-level planning cycles covering near-term deliverables and long-term bets
• Vet end-to-end scientific proposals for sound method, scale, and readiness to run in production
• Drive production readiness for models processing millions of daily decisions, including monitoring and on-call practices
• Hire, coach, and manage applied scientists, including performance management and promotion readiness

What it asks for

• 3+ years of scientist or machine learning engineer management experience
• Knowledge of ML, NLP, information retrieval, and analytics
• Experience with large-scale machine learning systems, including profiling, debugging, and performance and scalability
• Experience communicating with technical and non-technical audiences, including C-level executives
• Experience building ML models or algorithms for business applications (preferred)
• Experience delivering complex deep learning or computer vision software systems to customers (preferred)

Worth knowing

• Production on-call practices and rapid response for AI/ML systems

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Amazonnorth_east before you apply. If we got something wrong, report this listing.