Building a One-Stop Machine Learning Platform: Meituan's Turing Platform

Meituan’s Turing platform consolidates the entire delivery‑order workflow—from massive data ingestion and feature generation to model training, evaluation, deployment, real‑time prediction, and AB testing—into a single, end‑to‑end system that evolved from a minimal MVP into a fully platformized solution, addressing speed, accuracy, and engineering‑algorithm decoupling while planning deeper deep‑learning integration.

Meituan Technology Team
Meituan Technology Team
Meituan Technology Team
Building a One-Stop Machine Learning Platform: Meituan's Turing Platform

This article shares Meituan's experience in building a one-stop machine learning platform for their delivery business. As Meituan's daily order volume exceeded 30 million in July 2019, the delivery system faced challenges in making algorithms faster, better, and more accurate. The article describes the evolution of Meituan's machine learning platform from MVP to platformization, addressing issues like repeated development and inconsistent feature calculations.

The Turing platform aims to provide end-to-end services covering data processing, feature production, model training, evaluation, deployment, online prediction, and effect evaluation. Key components include:

- Offline training platform with visual interface for drag-and-drop model building

- Model management platform supporting local and remote deployment modes

- Offline feature platform for converting Hive table data to online features

- Real-time feature platform for collecting and processing online business data

- AB experiment platform with specialized methods for delivery scenarios

The article details technical implementations including MLDL (Machine Learning Definition Language) for bridging offline and online processing, feature grouping optimization for Redis storage, and time-slice based AB testing for delivery scenarios. Future plans include strengthening deep learning capabilities and further decoupling algorithms from engineering.

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

AB testingfeature engineeringDeep Learningreal-time featuresMachine Learning PlatformModel Managementdelivery system
Meituan Technology Team
Written by

Meituan Technology Team

Over 10,000 engineers powering China’s leading lifestyle services e‑commerce platform. Supporting hundreds of millions of consumers, millions of merchants across 2,000+ industries. This is the public channel for the tech teams behind Meituan, Dianping, Meituan Waimai, Meituan Select, and related services.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.