AI 應用工程師 【Python/Node.js】

Công việc được cập nhật khoảng 5 giờ
Nhà tuyển dụng hoạt động 1 ngày trước

Mô tả công việc

【職責範圍 Responsibilities】

  • 打造 Agentic 自動化工作流:基於我們核心的多模態 RAG 引擎,設計並部署多步驟的 AI 工作流 (Agentic workflows) 與決策路由邏輯,並與我們的核心 API 無縫互動。
  • 確保 AI 輸出的穩定與合規 (Guardrailing):設計嚴格的輸入/輸出防護機制,防範 Prompt Injection、過濾敏感個資 (PII),並確保 LLM 產出 100% 符合預期的資料格式 (如嚴格的 JSON Schema),以便安全地傳遞給下游系統。
  • 建構穩健的企業系統整合與 ETL:開發高容錯、安全的 API 連結,串接各大企業系統 (如 SAP, Oracle, NetSuite, QuickBooks, Snowflake),並維護可靠的 ETL 資料管道,將傳統企業數據同步至現代多模態或向量資料庫 (如 SurrealDB, Qdrant, ChromaDB)。
  • 設計 Human-in-the-Loop (HITL):針對關鍵的商業操作(如下單、寫入 ERP),設計並整合「人機協作審核」介面,確保系統的最終安全性與準確率。
  • 效能與成本優化:優化 LLM Prompt 表現、精準控管 API 成本,並針對脆弱的第三方 API 設計積極的錯誤處理 (Error handling) 與重試機制 (Retry logic)。


Yêu cầu

【我們希望你具備 Requirements】

  • 3 年以上軟體工程經驗,且具備將 LLM 應用落地到正式環境 (Production-grade) 的實戰經驗。
  • 熟悉 Python 與 TypeScript/JavaScript,能靈活穿梭於 AI 生態圈與網頁後端架構之間。
  • 熟悉打造多步驟 AI 代理 (Agent-based) 的框架與工具 (如 Agno/Phidata, LangChain, LangGraph, AutoGen 等)。
  • 對多模態 RAG 架構與向量資料庫 (如 ChromaDB, Qdrant) 有清晰的概念理解,能有效地與我們的核心基礎設施對接。
  • 具備企業級資安與合規意識:能與客戶的 IT 或資安團隊協作,確保系統部署符合嚴格的企業合規標準。

【有這些大加分 Nice to Haves】

  • 熟悉 LLM 輸出結構化與防護框架 (如 Guardrails AI, Instructor, Outlines 或 Pydantic Validation)。
  • 具備處理複雜 ETL 資料管道,或熟悉 AWS S3 等雲端儲存架構的經驗。
  • 熟悉新創環境的敏捷步調,具備強烈的 Ownership 與問題解決能力。

2
Yêu cầu 3 năm kinh nghiệm
1,200,000 ~ 1,800,000 TWD / năm
Liên kết lời mời cá nhân
Đây là liên kết giới thiệu cá nhân của bạn cho lời mời làm việc. Bạn sẽ nhận được thông báo qua email khi ai đó nộp đơn cho vị trí thông qua liên kết công việc của bạn.
Share this job
People who applied for this job also applied for

Về chúng tôi

【關於 Data Munger:讓 AI 真正在企業落地的幕後推手】

現在市面上充滿了各種 AI 聊天機器人和 Demo,但多數都無法真正在企業內部落地。為什麼?因為它們看不懂企業真實的商業邏輯,無法解讀複雜的報表,也無法在現有的工作流程中執行任務。

Data Munger 正在改變這件事。

我們致力於打造企業級 AI 的「資料脈絡層」(Data Context Layer)——把企業既有、龐雜且缺乏結構的數據,轉化為 AI 能理解並直接執行的狀態。我們不要求客戶打掉重練,而是直接無縫對接現有的資料庫與 ERP 系統,讓 AI 真正成為驅動企業營運的核心基礎設施。我們的技術目前已實際應用於財務對帳、供應鏈管理、醫療合規等場景。

🚀 為什麼你該加入我們?

  • 解決真實世界的硬問題 我們做的不是套殼 AI。在這裡,你將面對真實世界中「髒且複雜」的資料挑戰,運用 RAG、多模態處理等前沿技術,解決財務、供應鏈、醫療合規等硬核的商業痛點。
  • 巨大的個人影響力 我們是一個總部位於台灣的精實團隊。這裡沒有繁文縟節,你不會只是大公司裡的一顆螺絲釘——你的每一行程式碼、每一個架構決策,都將直接塑造產品的未來。
  • 打造真正有用的 AI 如果你厭倦了只為展示而做的專案,渴望看到自己打造的系統在企業核心營運中運轉、幫客戶抓出錯誤並提升效率——這裡就是你的舞台。

我們正在尋找喜歡動手解決問題、對AI產品有熱忱,且渴望與新創團隊一起快速成長的夥伴。



官網: https://datamunger.io/

【About Data Munger: Making Enterprise AI Actually Work】

The market right now is flooded with AI chatbots and flashy demos, but most of them never actually survive in a real enterprise environment. Why? Because they don't understand real business logic, they can't interpret complex reports, and they can't execute tasks within existing workflows.

Data Munger is changing that.

We are building the "Data Context Layer" for enterprise AI. We take a company's existing, massive, and unstructured data and transform it into a state that AI can actually understand and act upon. We don't ask clients to rip and replace their systems; we plug seamlessly into their existing databases and ERPs, turning AI into the core infrastructure that drives business operations. Our technology is already deployed in real-world scenarios like financial reconciliation, supply chain management, and healthcare compliance.

🚀 Why Join Us?

  • Solve Hard, Real-World Problems We aren't building "wrapper AI." Here, you will tackle the messy, complex reality of enterprise data. You’ll leverage cutting-edge tech like RAG and multi-modal processing to solve hardcore business pain points across finance, supply chain, and healthcare.
  • Massive Individual Impact We are a lean team based in Taiwan. There is no corporate red tape here, and you won't be just another cog in a big machine. Every line of code you write and every architectural decision you make will directly shape the future of our product.
  • Build AI That Actually Matters If you are tired of building projects just for demos and crave seeing your systems run in core enterprise operations—catching errors and driving real efficiency for clients—this is your stage.

We are looking for hands-on problem solvers who are passionate about AI products and eager to scale fast with a startup team.