🌏 全球远程工作机会汇总

在招 0 仅中国地区远程 0 👁 累计浏览 339 更新时间

高级智能体 AI 工程师(NDA 保密项目)

Senior Agentic AI Engineer NDA

全球可投 设计 预计有效期:2026-12-24

🏢 公司 / 雇主

GT

发布于 2026-11-09 · 首次收录 2026-11-16 · 最后确认在线 2026-11-17 · 👁 浏览 2 次

🛠 技能要求

Agentic AILLMPythonLangChainLangGraphPydantic AIOpenAI Agents SDKMCPSQLpandasDuckDBIbisAzure OpenAIDockerPlotly

📋 职位要求

  • 设计、构建和演进用于数据探索、分析和对话式洞察的智能体 AI 工作流。 Design, build, and evolve agentic AI workflows for data exploration, analysis, and conversational insights.
  • 开发并改进可扩展的智能体框架,包括编排、可复用工具、技能和对话管理。 Develop and improve an extensible agentic framework, including orchestration, reusable tools, skills, and conversation management.
  • 构建与应用服务和受治理数据源的安全集成,包括沙箱化代码执行和 AI 驱动的数据可视化。 Build secure integrations with application services and governed data sources, including sandboxed code execution and AI-powered data visualizations.
  • 设计可配置的解决方案,支持客户专属的提示词、工具、技能和数据定义,同时避免重复核心架构。 Design configurable solutions supporting client-specific prompts, tools, skills, and data definitions without duplicating the core architecture.
  • 开发 MCP 服务器和客户端,实现智能体与应用之间的通信和数据交换。 Develop MCP servers and clients to enable communication and data exchange between agents and applications.
  • 参与架构决策,以及现有基于 LangChain 的解决方案的演进与迁移。 Contribute to architectural decisions and the evolution and migration of the existing LangChain-based solution.
  • 实施 LLM 评估、测试、可观测性和安全控制,确保 AI 输出准确、可靠且有数据支撑。 Implement LLM evaluation, testing, observability, and security controls to ensure accurate, reliable, and data-grounded AI outputs.
  • 端到端负责功能,从技术设计与实现到生产发布和持续改进,并与跨职能团队协作。 Own features end to end, from technical design and implementation to production release and continuous improvement, collaborating with cross-functional teams.
  • 6 年以上软件工程、数据分析或相关领域经验,其中包含 2 年以上智能体 AI 工程经验。 6+ years of experience in software engineering, data analytics, or related fields, including 2+ years in agentic AI engineering.
  • 扎实的 Python 开发能力,具有构建、测试和维护生产级应用的经验。 Strong Python development skills, with experience building, testing, and maintaining production-grade applications.
  • 具有使用 Pydantic AI、LangGraph、LangChain 或 OpenAI Agents SDK 等框架构建智能体 AI 系统的实战经验。 Hands-on experience building agentic AI systems using frameworks such as Pydantic AI, LangGraph, LangChain, or OpenAI Agents SDK.
  • 深入理解智能体模式,包括工具调用、多步编排、上下文管理、护栏机制以及人机协同工作流。 Strong understanding of agentic patterns, including tool calling, multi-step orchestration, context management, guardrails, and human-in-the-loop workflows.

📄 职位描述(中文)

GT 成立于 2019 年,由一位前 Apple、Nest 和 Google 高管创立。GT 的使命是将全球顶尖人才与英国、美国、加拿大、德国和荷兰高增长公司提供的产品类职业机会连接起来。受客户委托,GT 正在寻找一位高级智能体 AI 工程师,该人选需有志于大规模解决复杂的匹配与搜索问题,将机器学习和算法应用于数亿条真实世界数据记录。 关于客户 我们的客户是一家领先的全球管理咨询公司,以攻克世界上一些最复杂的商业挑战而闻名。公司专注于战略、转型和绩效提升,与各行业的大型组织合作,创造持久的影响力。该公司持续被评为最佳雇主,将深厚的行业专业知识与协作、创新的文化相结合。其位于欧洲的中心枢纽在支持 EMEA 地区的运营方面发挥着关键作用,确保大规模的卓越与高效。 关于项目与职位 我们正在寻找一位高级 AI 工程师,帮助设计、构建和持续演进一个 AI 驱动的对话式分析平台,使企业能够探索数据、生成洞察和可视化,并做出数据驱动的决策。该产品帮助用户分析促销表现、回答复杂的分析问题、构建图表和表格,并利用 LLM、智能体工作流和受治理的客户数据生成基于证据的建议。团队正在开发一个可扩展、安全且易于扩展的智能体 AI 框架,可为不同客户进行定制,并通过 Model Context Protocol(MCP)等标准与其他产品集成。现有解决方案基于 LangChain 构建,并计划对该框架进行演进和迁移。这是一个注重实操的工程职位,同时包含大量架构和系统设计工作。你将参与智能体 AI 的完整生命周期,从设计智能体工作流、工具和集成,到实现评估框架、提升可靠性,以及交付可投入生产的 AI 功能。 职责 设计、构建和演进用于数据探索、分析和对话式洞察的智能体 AI 工作流。 开发并改进可扩展的智能体框架,包括编排、可复用工具、技能和对话管理。 构建与应用服务和受治理数据源的安全集成,包括沙箱化代码执行和 AI 驱动的数据可视化。 设计可配置的解决方案,支持客户专属的提示词、工具、技能和数据定义,同时避免重复核心架构。 开发 MCP 服务器和客户端,实现智能体与应用之间的通信和数据交换。 参与架构决策,以及现有基于 LangChain 的解决方案的演进与迁移。 实施 LLM 评估、测试、可观测性和安全控制,确保 AI 输出准确、可靠且有数据支撑。 端到端负责功能,从技术设计与实现到生产发布和持续改进,与跨职能团队协作。 必备知识、技能与经验 核心要求 6 年以上软件工程、数据分析或相关领域经验,其中包含 2 年以上智能体 AI 工程经验。 扎实的 Python 开发能力,具有构建、测试和维护生产级应用的经验。 具有使用 Pydantic AI、LangGraph、LangChain 或 OpenAI Agents SDK 等框架构建智能体 AI 系统的实战经验。 深入理解智能体模式,包括工具调用、多步编排、上下文管理、护栏机制以及人机协同工作流。 具有 MCP(Model Context Protocol)以及多智能体或跨应用集成的经验。 具有开发数据分析 AI 智能体的经验,包括数据集探索、代码生成与执行以及结果解读。 出色的系统设计与架构能力,包括构建可配置、可扩展的解决方案并评估技术权衡。 理解 AI 安全、评估与可观测性,包括安全数据访问、LLM 测试、链路追踪和性能监控。 强烈的责任心、沟通能力和跨职能协作能力,并有使用 AI 编程助手(Claude Code、Codex 或类似工具)的实际经验。 强烈优先 具有 Pydantic AI / Pydantic AI Harness 和 LangChain 的经验,尤其是框架之间的迁移。 熟悉用于数据分析的 SQL、pandas、DuckDB 或 Ibis。 具有 Azure OpenAI、Docker 以及 Plotly 或 Dash 等分析可视化工具的经验。 加分项 具有零售、定价、促销分析或类似领域的经验。 具有将面向客户的 AI 产品从原型交付到生产的经验。 熟悉 React 和 FastAPI。 面试流程 GT 招聘官面试 文化契合度面试 技术面试 终面 录用通知 为什么加入我们的客户? 加入一家全球知名的咨询公司,构建解决实际业务问题的创新 AI 产品。 帮助塑造一个不断演进的智能体 AI 平台的架构和技术方向。 构建对准确性、数据安全和可靠性要求极高的生产级 AI 解决方案。 使用现代 AI 技术和开发工具,包括 Pydantic AI、MCP、Claude Code、Cursor 和 OpenAI Codex。 与经验丰富的工程师和产品团队协作,工作环境鼓励实验、持续改进和技术自主。 申请时请提及词语 **UNCOMPLICATED** 并附上标签 RMjQwZTo0NTc6ZmYxMjozOWQ6NmQ0Nzo4MGE2Ojc1NzM6NjIzMw==,以表明你已完整阅读该职位描述(#RMjQwZTo0NTc6ZmYxMjozOWQ6NmQ0Nzo4MGE2Ojc1NzM6NjIzMw==)。这是一项用于减少垃圾申请的测试功能。公司可以搜索这些词语来找到完整阅读职位描述的申请者。
🇬🇧 查看英文原文(Original description)
GT was founded in 2019 by a former Apple, Nest, and Google executive.  GT’s mission is to connect the world’s best talent with product careers offered by high-growth companies in the UK, USA, Canada, Germany, and the Netherlands. On behalf of the client, GT is looking for a Senior Agentic AI Engineer who is interested in solving complex matching and search problems at scale, applying ML and algorithms to hundreds of millions of real-world data records.  About the Client Our client is a leading global management consultancy known for tackling some of the world’s most complex business challenges. With a focus on strategy, transformation, and performance improvement, the firm partners with major organizations across industries to drive lasting impact. Recognized consistently as a top workplace, it combines deep industry expertise with a collaborative, innovative culture. Its centralized European hub plays a key role in supporting operations across the EMEA region, ensuring excellence and efficiency at scale.  About the Project & Role We are looking for a Senior AI Engineer to help design, build, and evolve an AI-powered conversational analytics platform that enables businesses to explore their data, generate insights and visualizations, and make data-driven decisions. The product helps users analyze promotional performance, answer complex analytical questions, build charts and tables, and generate evidence-based recommendations using LLMs, agentic workflows, and governed client data. The team is developing a scalable, secure, and extensible agentic AI framework that can be customized for different clients and integrated with other products through standards such as Model Context Protocol (MCP). The existing solution is built on LangChain, with plans to evolve and migrate the framework. This is a hands-on engineering role with a significant architecture and system design component. You will work across the entire agentic AI lifecycle, from designing agent workflows, tools, and integrations to implementing evaluation frameworks, improving reliability, and delivering production-ready AI features. Responsibilities Design, build, and evolve agentic AI workflows for data exploration, analysis, and conversational insights. Develop and improve an extensible agentic framework , including orchestration, reusable tools, skills, and conversation management. Build secure integrations with application services and governed data sources, including sandboxed code execution and AI-powered data visualizations. Design configurable solutions supporting client-specific prompts, tools, skills, and data definitions without duplicating the core architecture. Develop MCP servers and clients to enable communication and data exchange between agents and applications. Contribute to architectural decisions and the evolution and migration of the existing LangChain-based solution . Implement LLM evaluation, testing, observability, and security controls to ensure accurate, reliable, and data-grounded AI outputs. Own features end to end, from technical design and implementation to production release and continuous improvement, collaborating with cross-functional teams. Essential Knowledge, Skills & Experience Core Requirements 6+ years of experience in software engineering, data analytics, or related fields, including 2+ years in agentic AI engineering . Strong Python development skills, with experience building, testing, and maintaining production-grade applications. Hands-on experience building agentic AI systems using frameworks such as Pydantic AI, LangGraph, LangChain, or OpenAI Agents SDK. Strong understanding of agentic patterns, including tool calling, multi-step orchestration, context management, guardrails, and human-in-the-loop workflows. Experience with MCP (Model Context Protocol) and multi-agent or cross-application integrations. Experience developing AI agents for data analysis, including dataset exploration, code generation and execution, and result interpretation. Strong system design and architecture skills, including building configurable, extensible solutions and evaluating technical trade-offs. Understanding of AI security, evaluation, and observability, including secure data access, LLM testing, tracing, and performance monitoring. Strong ownership, communication, and cross-functional collaboration skills, with practical experience using AI coding assistants (Claude Code, Codex, or similar). Strongly Preferred Experience with Pydantic AI / Pydantic AI Harness and LangChain , particularly migration between frameworks. Familiarity with SQL, pandas, DuckDB, or Ibis for data analysis. Experience with Azure OpenAI, Docker, and analytical visualization tools such as Plotly or Dash. Nice-to-Have Experience with retail, pricing, promotions analytics, or similar domains. Experience delivering client-facing AI products from prototype to production. Familiarity with React and FastAPI. Interview Process Interview with GT Recruiter Cultural Fit Interview Technical Interview Final Interview Offer Why Join Our Client? Join a globally recognized consulting firm building innovative AI-powered products that solve real business problems. Help shape the architecture and technical direction of an evolving agentic AI platform . Build production AI solutions where accuracy, data security, and reliability are essential. Work with modern AI technologies and development tools, including Pydantic AI, MCP, Claude Code, Cursor, and OpenAI Codex . Collaborate with experienced engineers and product teams in an environment that encourages experimentation, continuous improvement, and technical ownership. Please mention the word **UNCOMPLICATED** and tag RMjQwZTo0NTc6ZmYxMjozOWQ6NmQ0Nzo4MGE2Ojc1NzM6NjIzMw== when applying to show you read the job post completely (#RMjQwZTo0NTc6ZmYxMjozOWQ6NmQ0Nzo4MGE2Ojc1NzM6NjIzMw==). This is a beta feature to avoid spam applicants. Companies can search these words to find applicants that read this and see they're human.