The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Live from TWIMLcon! Use-Case Driven ML Platforms with Franziska Bell - #307
Oct 10, 2019
Franziska Bell, Ph.D., the Director of Data Science Platforms at Uber, shares cutting-edge insights on democratizing data science across the company. She discusses how use cases drive platform development, enhancing workflows for all employees. The collaboration with Uber's Michelangelo platform is highlighted, revealing challenges in integrating legacy models. Franziska also emphasizes the importance of open source in machine learning, innovative automation for data analytics, and fostering a collaborative culture within data science teams.
AI Snips
Chapters
Transcript
Episode notes
Democratizing Data Science
- Fran Bell's team aims to democratize data science within Uber, enabling anyone to access advanced tools.
- Their platforms, covering forecasting, anomaly detection, and conversational AI, empower users with push-button access to insights.
Conversational AI for Customer Support
- Uber uses conversational AI to improve customer support through its Customer Obsession Ticket Assistant.
- This tool aids representatives by recommending topics, actions, and responses, enhancing the customer care experience.
Platformization Criteria
- Consider three factors when platformizing data science capabilities: potential for step-function improvements, breadth of use cases, and module reusability.
- Fran Bell's team uses these criteria at Uber.