Algorithm-Practice-in-Industry is a curated resource collection for practitioners working on search, recommendation, and advertising algorithms in industry settings.
The project addresses the need to gather and organize practical knowledge from industry implementations of search, recommendation, advertising, and user growth systems. It collects articles from sources including Zhihu, Datafuntalk, and technical public accounts, organizing them into accessible reference materials. The core approach combines manual curation of industry practice articles with automated systems for tracking academic research: it maintains a spreadsheet of curated articles, generates daily summaries of new arXiv papers in information retrieval using large language models for ranking and translation, and aggregates papers from major conferences including ACL, CIKM, ECIR, EMNLP, ICLR, ICML, KDD, NAACL, NIPS, RecSys, SIGIR, WSDM, and WWW.
This resource suits practitioners and researchers focused on search, recommendation, and advertising who want both industry case studies and academic context. The project is particularly useful for those tracking recent developments across multiple top-tier conferences and seeking daily updates on new research. The tool automates paper discovery and summarization through arXiv monitoring and conference tracking, reducing manual effort in staying current with the field. Contributions are accepted through GitHub issues with provided templates, which trigger automated updates to the README and source data files.
The project maintains active development of its paper tracking pipeline with multiple entry points for different workflows including arXiv daily processing, conference paper aggregation, and industry practice maintenance. Automation is implemented through GitHub Actions workflows that handle scheduled arXiv updates, Feishu bot notifications, conference paper updates, and web page deployment. Legacy code has been archived separately, keeping the main development focused in the paperBotV2 directory while preserving historical scripts for reference and rollback purposes.
