Applied Researcher - Multi-Agent Systems
Core
Design, architect, and build robust multi-agent ecosystems and agentic workflows to power personalized recommendations at scale for eBay's global marketplace.
Role type
Senior Applied Researcher (Multi-Agent Systems & Agentic Workflows)
Builds
Autonomous multi-agent systems, agentic workflows, and next-generation recommendation experiences for e-commerce.
Domain
E-commerce, Large Language Models (LLMs), Multi-Agent Orchestration, Recommender Systems.
Deliverable
production ML models
Required skills
Multi-agent system design, agentic workflow orchestration, LLM/foundation model integration, NLP, recommender systems, Python, PyTorch, cloud services, big data pipelines, A/B testing, research publication record.
Preferred skills
Experience with ReAct patterns, tool-calling loops, planning systems, structured communication protocols, reflection loops, evaluator-generator frameworks.
Technologies
Python, PyTorch, LLMs, Foundation Models, Cloud Services, Big Data Pipelines.
Responsibilities
Design and evaluate multi-step agentic systems where agents coordinate to solve complex user intents; Investigate and implement state-of-the-art techniques in agent interaction patterns; Address frontier challenges in multi-agent production systems including latency, cost, and safety; Drive marketplace impact through rigorous A/B testing of autonomous user experiences.
Seniority
Senior, hands-on IC with strategic scientific direction
