{
    "schema_version": "via_ai_product_manager_jobs_v1",
    "data_type": "ai_product_manager_jobs",
    "job": {
        "job_title": "Agent应用工程师 (Agent Engineer)",
        "job_level": "初级",AI生成
        "work_category": "全职",AI生成
        "department": "未披露",
        "location": "广州·海珠区·凤阳",
        "work_mode": "现场办公",AI生成
        "job_category": "AI应用软件工程师"AI生成
    },
    "company": {
        "company_name": "心知引擎",
        "company_industry": "人工智能",
        "company_size": "20-99人",
        "company_website": "未披露",
        "company_description": "1. Company Overview\nDigitPulseAI is a technology company focused on large language model research and application development. Headquartered in Singapore, it has established branches in North America, Japan, Vietnam, and China. The company has raised tens of millions of USD in seed funding. Its vision is to leverage generative AI to create user value, with current primary business focus on AI-powered virtual social entertainment products.\n\n2. Team\n• Founder: Long  \n\n  Holds a Ph.D. from the University of Tokyo. Previously worked at leading tech companies including Oracle, Samsung, and Tencent, contributing to the R&D and product development of several social products with over 100 million daily active users (DAU).\n\n• Co-founder: Zhang Qi  \n\n  Ph.D. from the University of Toronto, supervised by Geoffrey Hinton. Senior deep learning researcher and expert in Mixture of Experts (MoE) systems.\n\n• Product Lead: Pei Liu  \n\n  Senior specialist in social and gaming communities. Formerly employed at prominent social entertainment companies like Tencent and Baiguoyuan, responsible for multiple products with tens of millions of DAU.\n\n• Expert Advisor: Michiko Asano  \n\n  Ph.D. in Psychology, Associate Professor at the University of Tokyo, and senior expert in linguistic and cognitive psychology.\n\n3. Business Focus\nThe company specializes in generative AI-driven social entertainment products. Key regional roles include:  \n• North America: Leads model training and core R&D.  \n\n• Singapore, Japan, Vietnam, and China: Primarily responsible for application development and implementation."
    },
    "compensation": {
        "salary_range": {
            "min": 20000,
            "max": 40000,
            "currency": "元",
            "cycle": "月"
        },
        "benefits": "法定节假日三薪 | 五险一金 | 加班补助 | 带薪年假 | 意外险 | 餐补 | 零食下午茶 | 绩效奖金 | 住房补贴 | 夜班补助 | 节假日加班费 | 年终奖 | 补充医疗保险 | 股票期权 | 企业年金 | 保底工资 | 定期体检 | 全勤奖 | 团建聚餐 | 底薪加提成"
    },
    "requirements": {
        "working_years": 1,
        "degree": "本科",
        "language": "中文",AI生成
        "visa_support": "未披露",
        "responsibilities": "•负责 Agent、Prompt、Memory、上下文策略优化\n•推进工具调用、策略切换、记忆检索等能力工程化\n•开发具备规划、执行、反馈闭环的自主Agent\n•实现工具学习(Tool Learning)与Function Calling\n•构建Agent评估体系与安全防护机制\n人才画像:\n•做过Agent、RAG、Memory 系统\n•1年+ LLM应用开发经验,精通Prompt Engineering\n•熟悉OpenAI Function Calling / Claude Tool Use等接口\n•有RAG、知识图谱与Agent结合的实际项目经验\n•有C端、生图、生视频、AI 陪伴、AI 社交经验更优",
        "requirements_text": "开发具备规划、执⾏、反馈闭环的自主Agent；1年+ LLM应用开发经验，精通Prompt Engineering；熟悉OpenAI Function Calling / Claude Tool Use等接口",
        "skills": [AI生成
            "Agent系统构建",
            "Prompt工程",
            "RAG与检索增强",
            "上下文/知识工程",
            "多模态AI应用",
            "大模型原理与能力边界",
            "工具调用/MCP",
            "编程与代码实现",
            "项目推动与执行闭环"
        ],
        "preferred_qualifications": "未披露",
        "job_highlights": "经验不限 | 本科"
    },
    "contact": {
        "recruiter_name": "雷先生",
        "contact_methods": [
            {
                "type": "apply_link",
                "value": "https://www.zhipin.com/job_detail/64be31538861d5550nV_0ty5FlJR.html"
            }
        ]
    },
    "posting": {
        "posting_id": "64be31538861d5550nV_0ty5FlJR",
        "post_date": "",
        "hire_time": "未披露",
        "update_date": "20260915",
        "job_url": "https://www.zhipin.com/job_detail/64be31538861d5550nV_0ty5FlJR.html"
    },
    "fulltext": "大模型算法\n多模态算法\nOpenAI\nC端\nPython\n核心职责：\n•负责 Agent、Prompt、Memory、上下文策略优化\n•推进工具调用、策略切换、记忆检索等能⼒⼯程化\n•开发具备规划、执⾏、反馈闭环的自主Agent\n•实现工具学习（Tool Learning）与Function Calling\n•构建Agent评估体系与安全防护机制\n⼈才画像：\n•做过Agent、RAG、Memory 系统\n•1年+ LLM应用开发经验，精通Prompt Engineering\n•熟悉OpenAI Function Calling / Claude Tool Use等接口\n•有RAG、知识图谱与Agent结合的实际项目经验\n•有C端、生图、生视频、AI 陪伴、AI 社交经验更优"
}