AI Engineering Architecture Salon: Backend Design for Intelligent Customer Service, Speech Recognition Engine, Lingxi Voice Analysis Platform, and High‑Performance Vector Retrieval

AI systems consist of algorithm models and engineering architectures; while deep learning has popularized the former, the latter remains essential for building complete AI solutions. To address the challenge of designing stable, flexible engineering architectures that support efficient algorithm iteration, 58.com AI Lab is hosting an AI Engineering Architecture Salon.

Event Information

Format: Online live stream Organizers: 58.com AI Lab, 58.com Technical Committee AI Branch, AICUG Producer: Zhan Kunlin, Head of 58.com AI Lab / Chair of AI Branch Dates: December 3 and December 8, 2020, 18:30‑20:00

Session 1 (Dec 3, 18:30‑19:15): Intelligent Customer Service Backend Architecture Abstract: Introduction of the "Bangbang" merchant‑version intelligent客服 platform, its micro‑service architecture, data storage, dialogue management, plug‑in QABot/TaskBot design, and practical results in the Yellow Pages micro‑chat operation scenario. Speaker: He Shuai, senior backend engineer at 58.com AI Lab.

Session 2 (Dec 3, 19:15‑20:00): Speech Recognition Engine Backend Architecture Abstract: Overview of the end‑to‑end speech recognition service, including audio file parsing, decoding, Docker deployment, real‑time streaming recognition, and optimization experiences. Speaker: Wang Yan, backend architect at 58.com AI Lab.

Session 3 (Dec 8, 18:30‑19:15): "Lingxi" Intelligent Voice Analysis Platform Backend Architecture Abstract: Presentation of the platform’s overall architecture, massive voice data storage, voice‑to‑text conversion, semantic tag extraction, and web visualization integration. Speaker: Liu Shengyuan, senior backend engineer at 58.com AI Lab.

Session 4 (Dec 8, 19:15‑20:00): High‑Performance Vector Retrieval Platform Abstract: Design of a Faiss‑based vector retrieval system covering incremental indexing, distributed indexing, retrieval services, and performance tuning. Speaker: Chen Zelong, senior backend engineer at 58.com AI Lab.

Additional Information: Recruitment opportunities at 58.com AI Lab and introductions to open‑source projects qa_match (a deep‑learning based QA matching tool) and dl_inference (a general deep‑learning inference service), with GitHub links provided.