高级软件AI工程师
Sr Software AI Engineer
客户是全球最大的健康、保健和美容产品电商零售商之一，为185多个国家的客户提供服务。他们拥有超过30,000种产品的目录和全球物流基础设施，每天帮助数百万人过上更健康的生活。他们正在围绕“AI优先”的战略扩展工程团队：将AI不仅作为产品功能，更是作为构建软件、服务客户和运营业务的基础。
他们正在寻找一位高级AI软件工程师，该职位需要编写生产代码，从规格到部署、运维和可观测性全程负责功能开发，并参与团队发布内容的值班轮班。没有单独的运维或可靠性职能进行交接；工程师负责整个生命周期。测试自动化通过AI驱动的工具集成到发布流程中，并使用黄金路径。
构建由GenAI驱动的产品体验以及支撑这些体验的共享AI平台基础设施。这包括针对目录和客户评论的RAG管道、LLM驱动的个性化推荐、对话式健康助手、代理工作流系统，以及使AI功能具备生产级和可重复性的评估和MLOps层。此方向的专业领域包括：RAG与个性化、代理框架与工具使用、评估与安全机制，以及面向内部业务功能（如营销自动化和BI代理）的LLM应用开发。
你将负责：
· 设计、构建和运行生产级AI功能：RAG管道、LLM驱动的推荐、对话式代理或代理工作流自动化。
· 构建共享的AI平台层：检索基础设施、评估框架、模型监控、安全机制和可观测性。
· 编写LLM应用并集成到营销平台、BI工具或面向客户的產品界面。
· 使用结构化的评估框架评估模型和特征质量；通过数据迭代提示、检索策略和模型选择。
· 在日常工作中使用AI驱动的SDLC工具，例如Claude Code，用于AI和非AI代码。
· 与个性化团队协作，确保GenAI产品功能与现有的ML个性化信号保持一致。
· 在共享知识库中记录AI系统设计决策、评估结果和运维经验。
· 负责你所构建的AI系统的可观测性：延迟、成本、质量漂移和错误率；参与值班轮班
查看英文原文
Client is one of the world's largest e-commerce retailers of health, wellness, and beauty products, serving customers in more than 185 countries. With a catalog of over 30,000 products and global logistics infrastructure, they help millions of people live healthier lives every day. They are growing its engineering organization around an AI-first mandate: using AI not just as a product feature but as the foundation for how they build software, serve customers, and operate the business.
They are looking for a Sr. Software AI Engineer who will write production code, own features end-to-end from spec through deployment, operations, and observability, and participate in on-call rotation for what the team ships. There is no handoff to a separate ops or reliability function; engineers own the full lifecycle. Test automation is built into the shipping process using AI-driven tooling and the golden path.
Build the GenAI-powered product experiences and the shared AI platform infrastructure that powers them. This includes RAG pipelines over the catalog and customer reviews, LLM-driven personalization, a conversational Wellness Agent, agentic workflow systems, and the evals and MLOps layer that makes AI features production-grade and repeatable. Specializations within this track include: RAG and personalization, agent framework and tool use, evals and guardrails, and LLM application development for internal business functions such as marketing automation and BI agents.
What you will do:
· Design, build, and operate production AI features: RAG pipelines, LLM-driven recommendations, conversational agents, or agentic workflow automation.
· Build the shared AI platform layer: retrieval infrastructure, eval frameworks, model monitoring, guardrails, and observability.
· Write LLM applications and integrations with marketing platforms, BI tools, or customer-facing product surfaces.
· Evaluate model and feature quality using structured eval frameworks; iterate on prompts, retrieval strategies, and model selection using data.
· Use AI-driven SDLC tooling such as Claude Code as a daily practice for both AI and non-AI code.
· Coordinate with the Personalization team to align GenAI product features with existing ML personalization signals.
· Document AI system design decisions, evaluation results, and operational lessons in the shared knowledge base.
· Own the observability of AI systems you build: latency, cost, quality drift, and error rates; participate in on-call rotation and respond to production incidents.
- 8+ years of software engineering experience. Fully autonomous; drives technical decisions within the team; mentors junior engineers.
- Python proficiency; comfortable building and operating production LLM applications.
- Hands-on experience with at least one specialization: RAG and retrieval systems, LLM evaluation, agentic frameworks (LangChain, LlamaIndex, or similar), or LLM-based workflow automation.
- Understanding of prompt engineering, context window management, and LLM output quality tradeoffs.
- Familiarity with vector databases, embedding models, or semantic search.
- AI-driven SDLC : hands-on experience shipping production code with AI-assisted development tools such as Claude Code, GitHub Copilot, or Cursor. The bar is not awareness; it is daily use in delivering real software.
- Full-stack awareness: comfortable contributing across layers of the stack when needed; purely single-layer specialists are not the target profile.
- Production ownership: experience owning features end-to-end from spec through deployment,
- NICE TO HAVE
- Exposure to MLOps tooling or model deployment pipelines.
- Contributions to internal developer tooling, golden path standards, or SDLC process improvements.
- Experience with e-commerce platforms, product catalogs, or high-traffic consumer applications.
- Exposure to MLOps tooling or model deployment pipelines.
- Experience working in distributed teams across different time zones / geographies.
- Track record of documenting architectural decisions, writing RFCs, or contributing to engineering wikis.
- Selected candidates will be invited to take part in several rounds of interviews.
- The role is expected to be full time and ideal candidates should be looking for a long term engagement.
- A background check will be required as part of the onboarding process.