Qunar Tech Salon is a learning and exchange platform for Qunar engineers and industry peers. We share cutting-edge technology trends and topics, providing a free platform for mid-to-senior technical professionals to exchange and learn.
This article introduces the OpenAI Agents SDK, explains its key capabilities such as Agent Loop, Handoffs, Guardrails, and Tracing, provides practical Python code examples, compares it with other multi‑agent frameworks, and discusses best practices for building reliable AI applications.
This technical case study details how a high‑traffic Kafka logging cluster was optimized by adjusting Filebeat and Kafka parameters, increasing compression batch size, and tuning Kubernetes settings, resulting in significant reductions in request volume, network traffic, CPU usage, and overall resource consumption.
This document describes a comprehensive capacity‑planning solution that predicts traffic‑peak impacts for hotel services, automatically estimates required CPU resources, creates timed scaling tasks, and evaluates performance using detailed metrics, thereby improving operational efficiency and reducing manual effort during events such as exam‑ticket printing and holiday travel surges.
The article details Qunar’s hotel search ranking system evolution, covering the shift from rule‑based sorting to LambdaMart, the adoption of LambdaDNN deep models, multi‑objective MMOE architectures, multi‑scenario integration, extensive feature engineering, and experimental results demonstrating significant offline and online performance gains.
In the AI era, Qunar's front‑end tech director presents how D2C and large‑model‑based code generation dramatically improve front‑end productivity across three business scenarios, achieving up to 55% automated code coverage for both C‑end and B‑end applications.
This article explains what constitutes a small file in HDFS, why they harm memory, compute and cluster load, outlines common sources such as data sources, streaming and dynamic partitioning, and provides detailed Hive and Spark solutions—including CombineHiveInputFormat, merge parameters, distribute by, and custom Spark extensions—to efficiently merge small files and improve job performance.
This document outlines the terminology, background, requirements, task classifications, state model, and detailed architecture—including TaskScheduler, TaskWorker, and TaskConsole components—of a new distributed task scheduling system designed to replace Celery in a database automation platform, with emphasis on scalability, reliability, and extensibility.
This article explains the necessity of vector databases for AI workloads, reviews the PostgreSQL ecosystem, compares vector database options, provides detailed PgVector installation and usage steps, shares operational best‑practices, performance tuning tips, and real‑world deployment cases at Qunar and Tujia.
This article examines various image‑generation approaches, compares web‑frontend, client‑side, and backend methods, introduces a new Node‑backend solution based on Satori to convert HTML to SVG and then to PNG with Resvg (later Sharp), and details performance and memory optimizations that dramatically improve speed, resource usage, and stability for large‑scale image‑service deployments.
At the WOT2024 conference, Qunar Travel’s CTO Zheng Jimin presented a comprehensive overview of AI-generated video production for international hotels, detailing challenges, AI-driven workflow automation, practical implementation steps, multilingual translation enhancements, and performance results, offering valuable insights for scaling high‑quality hotel video content.
