Author

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.

646

Articles

Likes

Views

Comments

Latest from Meituan Technology Team

100 recent articles max

Meituan Technology Team

Oct 17, 2024 · Frontend Development

How Recce Boosts Dynamic Front‑End Containers to Near‑Native Performance

This article analyzes Meituan's Recce solution for dynamic front‑end containers, detailing performance bottlenecks, architectural classifications, interpreter and language choices, UI framework design, rendering layer decisions, optimization techniques, and future directions to achieve native‑level speed while retaining dynamic capabilities.

RecceRustUI framework

0 likes · 28 min read

How Recce Boosts Dynamic Front‑End Containers to Near‑Native Performance

Meituan Technology Team

Oct 17, 2024 · Artificial Intelligence

Meituan Robotics Research Institute 2024 Call for Research Proposals

The Meituan Robotics Research Institute (MARS) is calling full‑time university scholars and researchers to submit independent research proposals for 2024 projects—selected from a predefined topic list, evaluated by Meituan and external experts on novelty, business value and feasibility, and eligible for up‑to ¥200,000 funding, on‑site interns, fast‑track graduate hiring and de‑identified data, with applications due 10 Nov 2024 and projects starting Dec 2024 or Jan 2025.

AIResearch FundingRobotics

0 likes · 4 min read

Meituan Robotics Research Institute 2024 Call for Research Proposals

Meituan Technology Team

Oct 10, 2024 · Artificial Intelligence

Global User Modeling and Explicit Interest Transfer Framework for Meituan Home Page Recommendation

Meituan’s home‑page recommendation system adopts a multi‑stage global user‑modeling pipeline culminating in the EXIT framework, which explicitly transfers cross‑domain interests via interest‑combination labels and a scene‑selector network, thereby mitigating data sparsity and negative transfer and delivering significant offline and online performance gains.

Meituancross-domain recommendationglobal user modeling

0 likes · 34 min read

Global User Modeling and Explicit Interest Transfer Framework for Meituan Home Page Recommendation

Meituan Technology Team

Sep 12, 2024 · Artificial Intelligence

How BlackPearl Dominated All Three KDD 2024 OAG‑Challenge Tracks with Large‑Model Techniques

The BlackPearl team from Meituan’s Search & Content Intelligence group detailed their award‑winning solutions for the three KDD 2024 OAG‑Challenge tasks—paper name disambiguation, source tracing, and academic QA—showcasing large‑model driven pipelines, iterative self‑refinement, grafting‑learning, and extensive hard‑negative mining that outperformed traditional feature‑engineered and BERT‑based baselines.

Academic Knowledge GraphInformation RetrievalKDD Cup

0 likes · 20 min read

How BlackPearl Dominated All Three KDD 2024 OAG‑Challenge Tracks with Large‑Model Techniques

Meituan Technology Team

Sep 5, 2024 · Big Data

Next‑Generation AB Experiment Analysis Engine for Multi‑Sided Scenarios

To overcome the limitations of traditional A/B engines in multi‑sided, small‑sample and spill‑over contexts, the article proposes a next‑generation analysis engine that standardizes adaptive workflows, automates method selection, integrates variance‑reduction and meta‑analysis techniques, and offers a modular, self‑service platform for robust, scalable experimentation.

AB testingExperiment analysismulti-sided experiments

0 likes · 23 min read

Next‑Generation AB Experiment Analysis Engine for Multi‑Sided Scenarios

Meituan Technology Team

Aug 15, 2024 · Artificial Intelligence

Meituan's Exploration and Practice in Advertising Algorithm: Information Flow Ad Estimation

Meituan’s advertising algorithm research, presented at Tech Salon #81, outlines the evolution of its information‑flow ad CTR estimation from tree models to sparse large‑scale DNNs, describes three current modeling directions—user‑side timelines and spatial lines, link‑side page/card reconstruction, and LLM‑based knowledge injection—and details practical implementations such as decision‑path modules, ultra‑long/ultra‑wide sequence handling, full‑reconstruction of pages and cards with a Context Modeling Transformer, concluding that combining algorithmic innovation with engineering effort is essential while large‑model integration remains a long‑term challenge.

AdvertisingCTRLLM

0 likes · 19 min read

Meituan's Exploration and Practice in Advertising Algorithm: Information Flow Ad Estimation

Meituan Technology Team

Aug 8, 2024 · Artificial Intelligence

BlackPearl Team Wins All Three Tracks of KDD 2024 OAG‑Challenge Cup with Large‑Model Solutions

The BlackPearl team from Meituan’s Dazhong Dianping division swept all three KDD 2024 OAG‑Challenge Cup tracks—WhoIsWho, PST, and AQA—by deploying innovative large‑model techniques such as iterative text clustering, graft‑learning‑enhanced BERT RAG pipelines, and a Boosting LLM‑for‑Vector search, and have released the code publicly on GitHub.

Academic DisambiguationKDD CupLarge Language Model

0 likes · 4 min read

BlackPearl Team Wins All Three Tracks of KDD 2024 OAG‑Challenge Cup with Large‑Model Solutions

Meituan Technology Team

Aug 8, 2024 · Artificial Intelligence

Highlights of Meituan's ACL 2024 Papers: Speculative Decoding, Graph‑Structured Decoding, DolphCoder, and Instruction Fine‑tuning

Meituan showcases four ACL 2024 papers—Early‑Exiting Speculative Decoding with a Thompson‑sampling controller, Graph‑Structured Speculative Decoding that merges draft hypotheses in a DAG, DolphCoder, a code‑generation LLM improved by diverse multi‑objective instruction tuning, and a study of instruction fine‑tuning that finds it mainly aligns existing knowledge—while inviting attendees to its booth 11 and a live paper discussion on August 12.

ACLInstruction TuningLLM

0 likes · 8 min read

Highlights of Meituan's ACL 2024 Papers: Speculative Decoding, Graph‑Structured Decoding, DolphCoder, and Instruction Fine‑tuning

Meituan Technology Team

Jul 25, 2024 · Artificial Intelligence

Selected Meituan Papers Accepted at KDD 2024: Summaries of Five Long Papers

Meituan’s five long papers accepted at KDD 2024 introduce a dual‑intent model for search‑recommendation, a joint auction mechanism for ads, a robust ATE estimator for heavy‑tailed metrics, a decision‑focused causal learning framework for marketing, and an efficient on‑demand order‑pooling system for real‑time courier assignments.

Controlled ExperimentsKDD 2024Recommendation Systems

0 likes · 12 min read

Selected Meituan Papers Accepted at KDD 2024: Summaries of Five Long Papers

Meituan Technology Team

Jul 18, 2024 · Fundamentals

Multithreading Programming: Concepts, Synchronization, and Best Practices

Multithreaded programming splits tasks across logical and hardware threads to exploit multicore CPUs, requiring careful use of synchronization primitives such as mutexes, read‑write locks, condition variables, and lock‑free atomics, while avoiding pitfalls like race conditions, deadlocks, and false sharing for correct, high‑performance software.

CLockatomic

0 likes · 65 min read

Multithreading Programming: Concepts, Synchronization, and Best Practices