Visual Language Model and Agentic AI R&D Engineer(資通所_視覺語言模型與自主式智慧代理研發工程師F103)

【Job Description工作內容】

1. Develop, integrate, and validate Vision-Language Models (VLMs), Large Language Models (LLMs), and Agentic AI systems.

2. Build multimodal inference pipelines that combine images, text, and documents, and develop intelligent agent systems with task planning, tool use, and workflow automation capabilities.

3. Design prompts, RAG pipelines, tool-use mechanisms, workflow orchestration, and evaluation frameworks to improve system accuracy, robustness, and usability.

4. Collaborate with platform, infrastructure, frontend/backend, and application teams to bring generative AI technologies into real industrial scenarios.

【Position Highlights / Advantages】

1. Work on cutting-edge technologies including VLMs, LLMs, RAG, and AI agents.

2. Participate in building practical multimodal and agentic AI prototypes from proof-of-concept to deployable systems.

3. Gain end-to-end experience across cloud, edge, data engineering, and real-world application integration.

1. 負責視覺語言模型（VLM）、大型語言模型（LLM）與 Agentic AI 系統之研發、整合與驗證。

2. 建構多模態推論流程，整合影像、文字、文件等資料，發展具任務規劃、工具調用與流程自動化能力之智慧代理系統。

3. 設計 Prompt、RAG、Tool Use、Workflow Orchestration 與模型評測機制，提升系統準確度、穩定性與可用性。

4. 與平台、Infra、前後端及場域應用團隊合作，推動生成式 AI 技術於實際產業場域落地。

《職務亮點/優勢》

1. 可參與 VLM、LLM、RAG、AI Agent 等前瞻技術之系統化研發。

2. 有機會將多模態模型與 Agent workflow 導入真實產業情境，從 PoC 推進到可落地原型。

3. 可結合雲端、邊緣、資料工程與應用場域需求，累積完整生成式 AI 系統實作經驗。

【Job Requirements資格條件】

1. Master’s degree or above in Computer Science, Electrical Engineering, Information Management, Artificial Intelligence, or a related field.

2. Strong proficiency in Python, with experience in API integration, modular development, and debugging.

3. Familiarity with generative AI area, such as LLMs, VLMs, RAG, prompt engineering, function calling, or agent frameworks.

4. Experience with Hugging Face, Transformers, vLLM, LangChain, LlamaIndex, OpenAI API, or related open-source frameworks is a plus.

5. Familiarity with Linux, Docker, Git, and basic cloud/service deployment workflows is preferred.

【Preferred Qualifications】

1. Experience in multimodal model integration, document understanding, image understanding, or AI agent workflow design.

2. Experience in model evaluation, performance optimization, inference deployment, or MCP / tool-calling implementation.

3. Candidates with relevant projects, publications, GitHub repositories, or technical portfolios are encouraged to provide references.

1. 碩士(含)以上，資訊工程、電機、資管、人工智慧或相關科系畢。

2. 熟悉 Python 程式開發，具備 API 串接、模組化開發與除錯能力。

3. 熟悉相關生成式 AI 技術，如 LLM、VLM、RAG、Prompt Engineering、Function Calling 或 Agent Framework。

4. 具備 Hugging Face、Transformers、vLLM、LangChain、LlamaIndex、OpenAI API 或相關開源框架使用經驗者佳。

5. 熟悉 Linux、Docker、Git 與基本雲端/服務部署流程者佳。

《加分條件》

1. 具多模態模型整合、文件理解、影像理解或 AI Agent 工作流程設計經驗。

2. 具模型評測、效能優化、推論部署或 MCP / Tool Calling 實作經驗。

3. 若有相關專案成果、論文、GitHub 或技術作品，歡迎提供參考。

【Work Location工作地點】Hsinchu County, Taiwan新竹縣竹東鎮

【About Information and Communications Research Laboratories關於資通所】

Founded in 1990, the Information and Communications Research Laboratories (ICL) of ITRI focuses on the development of advanced and core technologies for the ICT industrial upgrading and value-added in Taiwan. From personal computers (including laptops) and network communications in the early days, to new smart devices and cloud application services, ICL built the foundation for industrial development by leading technology research and cultivating talents. In line with government policy, ICL also consolidates national industry, government and academic resources to participate in international ICT standard organizations to enhance Taiwan's visibility and influence in the global market.

In response to the cross-domain integration trend in the ICT industry, ICL combines the technical fields of communication, information and data science and focuses on the following four priorities, with the vision of developing essential ICT technologies and building Taiwan into a reliable global supply chain partner.

1. Establish Taiwan’s Core Technologies for the 5G/B5G/6G Industry
2. Develop National Satellite Payload and Ground Station Technologies
3. Integrate AI and ICT Technologies to Accelerate companies’ Digital Optimization and Transformation
4. Build a trustworthy ICT Cyber-Security System and Supply Chain worldwide

Looking forward to the future, ICL will accelerate the development of three application fields: smart life, healthy living, and sustainable environment, and by taking digital transformation as the guidance, to provide services on industry essential technology independence, embedded supply chain, and enterprise flexibility and resilience.

工研院資訊與通訊研究所(以下簡稱資通所)於1990年成立，以ICT產業發展所需之前瞻主題研究、關鍵技術開發、以其研發成果附加價值的提升與產業化為核心。從早期的個人電腦（含筆記型電腦）、有線通訊與網路、無線與行動通訊，一直到新興智慧終端與雲端物聯網應用服務等，資通所不僅僅是ICT技術的先驅領航者，也培育出相當豐沛且深厚的技術能量及優秀人才，為國內產業發展奠定殷實的根基。此外，資通所亦配合政府政策，從無到有地協助整合國內產官學研各界能量參與國際ICT相關標準組織活動與標準制定，藉由技術創新與智財的累積，更增進我國在全球競爭舞台的能見度與影響力。

因應當前ICT產業跨域整合趨勢，資通所結合通訊、資訊、資料科學三大技術領域且聚焦下列四項優先任務，以發展資通訊關鍵技術，打造台灣成為全球可信賴之供應鏈夥伴為願景。

• 建立台灣產業5G/B5G/6G優勢自主核心技術
• 發展國產衛星通訊酬載與地面系統技術能量
• 整合AI+資通訊技術加快企業數位優化與轉型
• 打造被世界信賴的ICT資安系統及供應鏈

展望未來，資通所將加速投入智慧生活、健康樂活、永續環境三大應用領域的發展，並以數位轉型為導向，提供產業關鍵技術自主、嵌入供應鏈、企業彈性與韌性的需求服務。