Predicting order batchability
summary
Describes why and when orders should be batched in online food delivery and the customer-experience and business impacts (churn, repeat, NPS) of failing delivery time promises. Explains Swiggy’s deployment of a deep learning system called FOPS to predict which orders will be batched, outlines problems in the prior framework and the operational, ML and engineering approaches used, and summarizes results from first-order batching prediction and planned next steps.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:24Overview
- 5:01Introduction
- 6:29Perks of Batching
- 11:35Next Version of Order Flow
- 12:43Second Order Batching Prediction
- 14:54Demand prediction
- 16:47Batching conditions
- 18:23Order Level Predictions
- 19:46Historical Features
- 22:14Experiment
related talks
41mCapacity and efficiency problems in real-time logistics - and how data at Swiggy solves this
18mBatchy: Batch scheduling Data Flow Graphs with Service level Objectives
27mOur experiments with food recommendations
26mDelivering Delivery Time Prediction
27mOptimizing Omnichannel Order Fulfillment with AI and Advanced Analytics
22mReal-world applications of machine learning: Beyond the hype
22mData-Driven F&B Delivery: Jahez as a Leading Example
29mBatch and Streaming Processing in the World of Data Engineering and Data Science