The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) How Machine Learning Powers On-Demand Logistics at Doordash with Gary Ren - #405
Aug 31, 2020
Gary Ren, a machine learning engineer at DoorDash, dives into the transformative role of machine learning in logistics. He shares insights on optimizing route planning and balancing the three-sided marketplace of consumers, dashers, and merchants. The conversation highlights the integration of predictive modeling for delivery timings and the innovative use of reinforcement learning to boost efficiency. Plus, Ren discusses challenges in unpredictable environments, including how to adapt to real-time conditions to improve customer satisfaction.
AI Snips
Chapters
Transcript
Episode notes
Restaurant Wait Time Example
- Sam Charrington mentions a Jamaican restaurant notorious for inaccurate wait times.
- This highlights the challenge of predicting food preparation time, a key aspect of DoorDash's logistics.
DoorDash Logistics Engine
- Gary Ren works on the DoorDash logistics team, which manages delivery fulfillment.
- This core engine includes balancing supply and demand, route planning, and delivery assignments.
Balancing Supply and Demand
- DoorDash balances supply (dashers) and demand (deliveries) using forecasting and real-time adjustments.
- They use dynamic pricing and messaging to influence both dasher availability and consumer demand.