Applied Scientist II, Demand Enablement, Product Analytics and Operations
Core
Design and build intelligent multi-agent systems that automate root cause analysis for advertising campaign delivery at scale.
Role type
Senior Applied Scientist (LLM-based agent architectures)
Builds
Production diagnostic systems for Amazon DSP
Domain
Programmatic advertising, distributed systems observability, LLM-based reasoning
Deliverable
production ML models
Required skills
LLM-based agent architectures, retrieval-augmented generation, time-series anomaly detection, causal inference, numerical optimization, parallel and distributed computing, statistical analysis
Preferred skills
deep learning algorithms, computer vision, professional software development, experimental design
Technologies
Java, C++, Python, foundation models
Responsibilities
Architect agentic orchestration patterns for specialized sub-agents, develop hierarchical analysis frameworks for trend detection and anomaly isolation, build self-learning feedback loops for diagnostic knowledge, collaborate with engineers to deliver end-to-end production solutions, run A/B experiments to optimize advertiser experiences, research and implement new ML models for advertising performance
Seniority
Senior, hands-on IC
