Sharing Zhihu tech posts and exploring community technology innovations.
Zhihu migrated thousands of Spark SQL jobs to the open‑source Gluten/Velox engine, achieving up to 2.7× speedup, 44% CPU savings and 57% memory reduction while fixing numerous consistency, performance, and stability issues through custom optimizer rules and targeted patches.
The project consolidates fragmented user data across departments into a unified OneID system, builds a comprehensive tagging and feature platform, and leverages real‑time and offline signals for algorithmic push, custom models, and crowd selection, enabling precise, data‑driven growth initiatives.
Zhihu created a unified OneID framework that merges scattered account, device, and behavior data into a global unique identifier, using strong and weak IDs, graph‑based connectivity, device governance, and a device half‑life model to improve recommendation, push, and advertising effectiveness.
This article details a multi‑year, data‑driven transformation in which a product‑research team leveraged large‑model AI and agentic workflows to automate repetitive coding, streamline hot‑topic discussion creation, and replace a seven‑person outsourcing crew, achieving up to 38.6% project‑time reduction, a 22.5‑25 PD weekly capacity gain, and a dramatic drop in marginal costs.
The article details the design of a cross‑platform C++ rendering engine built for a mobile reading app, explaining its three‑layer architecture, file‑format support, layout and caching mechanisms, security features, and how these technical choices delivered measurable improvements in page views, reading time, and conversion rates.
This article explores how multimodal large models like Qwen2.5‑VL enable high‑dimensional tag generation and universal embeddings for recommendation systems, detailing data synthesis, model training, quantization, fine‑tuning, and the resulting improvements in click‑through rate and exposure interaction.
This tutorial walks through using Zhihu's open‑source FUST microservice framework together with DeepSeek's language model API to design, implement, and deploy a scalable, high‑performance intelligent Q&A system, covering architecture, data models, service layers, and deployment scripts.
Zhihu partnered with Huawei to develop and open‑source an Appium driver for HarmonyOS, enabling unified, efficient automated testing across iOS, Android, and HarmonyOS and energizing the open‑source HarmonyOS ecosystem for developers and enterprises.
This article introduces ZHPluginSDK, a powerful Swift‑based plugin framework that supports type‑safe registration, flexible message dispatch, multi‑container composition, synchronous messaging, and environment variables, providing a comprehensive guide with architecture diagrams, feature lists, and extensive code examples for mobile developers.
This article introduces the open‑source Zhi‑Create‑Qwen3‑32B model, detailing its fine‑tuned training on creative‑writing data, the multi‑domain dataset strategy, curriculum‑learning based SFT, evaluation on WritingBench, and practical deployment options across various hardware and inference frameworks.
