Tianyu Chen
Senior Product Owner, Machine Learning Platform, eBay
He is responsible for the planning and construction of eBay's internal feature engineering platform, model training platform, model inference platform, and large model inference pipeline, playing the role of core pusher and enabler for the company's AI transformation strategy. Master of Software Engineering (Natural Language Processing Direction) from Fudan University, with rich experience in AI platform/application product R&D. He used to be an AI architect in Microsoft (Asia) Internet Engineering Institute, and has participated in and led the construction process of AI platforms for domestic and international head enterprises in finance, healthcare, automotive, real estate and other fields for many times.
Topic
GenAI-powered xGen e-commerce AI platform Product practices and reflections
eBay, a pioneer in the field of e-commerce, is about to celebrate a major milestone with its 30th anniversary. In today's eBay marketplace, billions of items flow between millions of active buyers and sellers. Generative Artificial Intelligence (GenAI) brings unprecedented scenarios and opportunities to this traditional e-commerce ecosystem, leading eBay to create unique user experiences and set new benchmarks in the industry that will shape the company's future. On the basis of eBay's self-research AI platform that has been precipitated for many years, in the face of the new paradigm and challenges brought about by the end-to-end product construction of GenAI, how do we build new platform capabilities and product experience, empower large model distributed development, achieve large planning and highly scalable large model reasoning, and accelerate the construction and implementation of GenAI applications, which has become the core driving force of the company's AI transformation. In this talk, I will provide insights into key product strategies and innovation practices for next-generation AI platforms, as well as constructive thoughts on key areas of focus for different companies' investments in generative AI. Outline (TBD): 1.GenAI-driven new practices for e-commerce platforms 2.New paradigm of GenAI application development 3.Key Issues and Challenges of AI Platforms 4. Core product path of AI platforms a)1 hour validation: GenAI Sandbox b)1 day optimisation: Zero-code large model fine-tuning c) 1 week on-line: RAG solution rapid construction 5. Platform Governance for Large Models a)Global GPU utilisation optimisation b)Responsible AI governance for all links 6.Extended Thinking a)GenAI's strategic investment in different enterprises b)Future Outlook
