資料工程師

【角色定位與任務】

你將建立安全、可追溯、可重複使用的AI-ready資料底座，支援RAG、模型訓練、分析與實際AI工作流程落地。這個角色適合願意面對真實業務複雜度，並希望讓AI能力真正進入營運場景的人。

我們重視資料工程基礎、快速學習、新挑戰意願，以及在安全與治理要求下打造可靠系統的能力。

【 主要負責工作 】

• 資料流處理：建置批次與即時pipeline，整合結構化、非結構化與多模態資料，支援向量檢索、RAG與AI應用

資料治理與品管：管理metadata、可追溯性、版本與品質規則，建立監控、異常處理與重跑機制，提升資料可信度

- 跨團隊協作與標準制定：與AI/ML、架構與業務團隊合作，定義資料介面、效能、安全與部署標準

- 資安法遵與生態系整合：落實權限控管、去識別化與個資法遵，串接跨系統與集團資料，記錄用途、限制與責任歸屬

【 相關經驗(符合 3 項以上歡迎投件) 】

- 實務資料工程經驗，能獨立設計及除錯可維運的ETL／ELT資料管線

- 具備至少一種資料處理、工作流編排或串流技術的實作經驗

- 理解資料模型、API、檔案／物件儲存、資料品質與效能最佳化

- 具處理文件、文字、影像、音訊或其他非結構化／多模態資料的經驗

- 能以清楚文件、測試及監控確保資料管線可被團隊長期維護

- 能與不同職能夥伴溝通資料限制、風險及取捨

【 技術能力 】

- Python、SQL

- ETL／ELT 資料管線的設計、除錯與效能調校

- 關聯式資料庫、檔案／物件儲存與資料建模

- 非結構化與多模態資料處理（文件、文字、影像或音訊）

- Git 版本控制，以及測試、文件與監控的工作習慣

- Linux 環境基本操作

【加分條件】

- 具向量資料庫、圖資料庫、搜尋引擎、embedding pipeline、RAG資料、多模態資料處理經驗。

- 熟悉metadata管理、資料血緣、資料目錄、資料契約或主資料管理。

- 具雲端或地端容器化環境、CI/CD及Infrastructure as Code經驗。

- 具批次或串流資料處理與工作流排程經驗（如 Airflow、Spark、Kafka 或同類工具）。

- 具金融、保險、電信、醫療或其他高監管產業經驗。

- 了解台灣個人資料保護法、資料去識別化及權限治理實務。

【 語言能力 】

- 中文：聽(精通)、說(精通)、讀(精通)、寫(精通)

- 英文：聽(中等)、說(中等)、讀(中上)、寫(中上)

【 到職後的成功樣貌 】

你能建立具明確owner、品質指標與監控的資料產品，讓團隊更快取得可信資料、更穩定地推出新AI場景，並在創新速度、資料治理與營運風險之間取得良好平衡。你的成果將成為人工智慧研發處持續擴展企業級AI能力的重要基礎。

※English Version

【 Role Overview and Mission 】

You will build a secure, traceable, and reusable AI-ready data foundation to support RAG, model training, analytics, and real-world AI workflow implementation. This role is ideal for someone who is motivated by real business complexity and wants to help bring AI capabilities into operational environments.

We value strong data engineering fundamentals, fast learning, willingness to take on new challenges, and the ability to build reliable systems under security and governance requirements.

【 Key Responsibilities 】

- Data pipeline processing: Build batch and real-time pipelines that integrate structured, unstructured, and multimodal data to support vector search, RAG, and AI applications.

- Data governance and quality control: Manage metadata, traceability, versioning, and quality rules; establish monitoring, exception handling, and rerun mechanisms to improve data reliability.

- Cross-functional collaboration and standards: Work with AI/ML, architecture, and business teams to define data interfaces, performance, security, and deployment standards.

- Security, compliance, and ecosystem integration: Implement access control, de-identification, and personal data compliance practices; connect data across systems and group entities while documenting usage, restrictions, and ownership.

【 Relevant Experience (Candidates who meet three or more are encouraged to apply) 】

- Hands-on data engineering experience, with the ability to independently design and debug maintainable ETL/ELT data pipelines.

- Practical experience with at least one data processing, workflow orchestration, or streaming technology.

- Understanding of data models, APIs, file/object storage, data quality, and performance optimization.

- Experience processing documents, text, images, audio, or other unstructured/multimodal data.

- Ability to ensure long-term maintainability of data pipelines through clear documentation, testing, and monitoring.

- Ability to communicate data limitations, risks, and trade-offs with partners across different functions.

【 Technical Skills 】

- Python and SQL

- Design, debugging, and performance tuning of ETL/ELT data pipelines

- Relational databases, file/object storage, and data modeling

- Unstructured and multimodal data processing, including documents, text, images, or audio

- Git version control, with strong habits in testing, documentation, and monitoring

- Basic Linux environment operations

【 Preferred Qualifications 】

- Experience with vector databases, graph databases, search engines, embedding pipelines, RAG data, or multimodal data processing.

- Familiarity with metadata management, data lineage, data catalogs, data contracts, or master data management.

- Experience with cloud or on-premises containerized environments, CI/CD, and Infrastructure as Code.

- Experience with batch or streaming data processing and workflow scheduling, such as Airflow, Spark, Kafka, or similar tools.

- Experience in finance, insurance, telecommunications, healthcare, or other highly regulated industries.

- Understanding of Taiwan’s Personal Data Protection Act, data de-identification, and access governance practices.

【 Language Requirements 】

- Mandarin Chinese: Proficient in listening, speaking, reading, and writing

- English: Intermediate listening and speaking; upper-intermediate reading and writing

【 What Success Looks Like 】

You will build data products with clear ownership, quality metrics, and monitoring, enabling the team to access trusted data faster and launch new AI use cases more reliably. You will help balance innovation speed, data governance, and operational risk. Your work will become a key foundation for the AI R&D division to continue scaling enterprise-grade AI capabilities.