Senior Software Engineer, Applied AI

免费注册后解锁"你的匹配分 vs 高分画像"与个人技能缺口。

About the role VIVIO Health is transforming how it operates with AI, and we are hiring a senior engineer to build the features that make that real: language-model capabilities for our clinical, client success, and member experience teams, built on AWS Bedrock and related models and shipped to real users. Your job is to take a business problem from one of those teams, work out where a model helps and where it does not, and deliver something the team uses every day. This is not a research role. The AI work lives inside our production platform, so you will also ship application features, work with the integrations that exchange data with our partners and build the test and quality practices that keep all of it reliable. Our clinical programs are the first area we are transforming with AI, so your early work will be with the clinical team and the documents, reviews, and decisions that run through it. Other business areas follow as the platform stabilizes. The role is remote, with structured written communication as the default working mode. You report to the Head of Engineering and work directly with the business owners of the problems you solve. What you will do Build and extend LLM-backed capabilities on AWS Bedrock and related models: document OCR and extraction on clinical and claims paperwork, summarization and decision support for clinical review, Knowledge Bases for retrieval over plan and policy documents, and workflow automation for internal teams Establish and maintain evals for the AI features you ship, so we know when quality drifts Handle the practical side of production AI: structured output, failure modes, cost, latency, and knowing when a deterministic approach beats a model Use agentic coding tools (Claude Code and Kiro) as a core part of your workflow, and own the verification of what they produce Design, build, test, and ship features in our Python and FastAPI back end on MySQL and AWS, and our Jinja2, Tailwind, and Alpine.js front end, with a bias toward fixing what is broken before adding what is new Work with the integrations that exchange eligibility, accumulator, and claim data with carriers and pharmacies, including NCPDP D.0 and X12 270/271 transactions and the file formats behind them Put test and lint gates in place where they are missing, and treat CI as part of the definition of done Treat PHI handling as a first-class design constraint in everything you build Write clear specs, prompts, and design notes; at this level, written communication is an engineering skill How we work Development happens in a managed Linux environment that you reach by remote desktop from a Windows endpoint; PHI never lands on a laptop. You need to be fluent in both: Linux for the actual engineering, Windows as the workstation you get there from. What we need 7+ years building production software, with real ownership of systems that other people depend on Strong Python with production FastAPI (or a comparable async framework) experience, solid SQL on MySQL or Postgres, and working knowledge of AWS Comfortable in a server-rendered front end (Jinja2, Tailwind, Alpine.js or htmx); we are not looking for a React specialist Fluent in Linux for development and comfortable working from a Windows workstation over remote desktop into a secure environment Daily use of AI coding tools on real work, and a clear point of view on where they help and where they fail At least one LLM-powered feature shipped to real users, including how you evaluated it and what went wrong Practical knowledge of prompt and context engineering, retrieval, embeddings, and structured output Working knowledge of healthcare data exchange standards and file-based integrations: NCPDP D.0, X12 270/271, SFTP, fixed-width and delimited formats, and reconciliation Direct communicator who names tradeoffs plainly and asks when something is unclear Nice to have Healthcare or other regulated-data experience (HIPAA, PHI, SOC 2) Pharmacy benefits, claims, or clinical program

关于该职位 VIVIO Health 正在通过 AI 改变其运营方式，我们正在招聘一名高级工程师来构建让这一切成为现实的功能：为我们的临床、客户成功和会员体验团队提供语言模型能力，基于 AWS Bedrock 及相关模型构建，并交付给真实用户。你的工作是接过这些团队中的一个业务问题，弄清楚模型在哪些地方有帮助、哪些地方没有，并交付团队每天都会使用的东西。这不是一个研究岗位。AI 工作存在于我们的生产平台内部，因此你还需要交付应用功能、处理与合作伙伴交换数据的集成，并建立测试和质量实践来确保这一切的可靠性。我们的临床项目是我们用 AI 转型的第一个领域，因此你的早期工作将与临床团队以及贯穿其中的文档、审核和决策打交道。随着平台趋于稳定，其他业务领域将陆续跟进。该职位为远程工作，以结构化的书面沟通为默认工作模式。你向工程负责人汇报，并直接与你所解决问题的业务负责人合作。 你将做什么 在 AWS Bedrock 及相关模型上构建和扩展由 LLM 支持的能力：临床和理赔文书上的文档 OCR 与提取、临床审核的摘要与决策支持、用于计划和政策文档检索的 Knowledge Bases，以及面向内部团队的工作流自动化 为你交付的 AI 功能建立并维护评估，以便我们了解质量何时发生漂移 处理生产 AI 的实际问题：结构化输出、故障模式、成本、延迟，以及知道何时确定性方法优于模型 将智能体编码工具（Claude Code 和 Kiro）作为工作流的核心部分，并对它们产出的内容负责验证 在我们的 Python 和 FastAPI 后端（基于 MySQL 和 AWS）以及我们的 Jinja2、Tailwind 和 Alpine.js 前端中设计、构建、测试和交付功能，并倾向于先修复已损坏的东西再添加新东西 处理与承运商和药房交换资格、累计额和理赔数据的集成，包括 NCPDP D.0 和 X12 270/271 交易及其背后的文件格式 在缺失的地方设置测试和 lint 门禁，并将 CI 视为完成定义的一部分 将 PHI 处理视为你所构建的一切的一等设计约束 编写清晰的规格说明、提示词和设计文档；在这个级别，书面沟通是一项工程技能 我们如何工作 开发在受管理的 Linux 环境中进行，你通过远程桌面从 Windows 端点访问；PHI 永远不会落在笔记本电脑上。你需要熟练掌握两者：Linux 用于实际的工程工作，Windows 作为你访问它的工作站。 我们需要什么 7 年以上构建生产软件的经验，并对他人依赖的系统有真正的所有权 扎实的 Python 和生产级 FastAPI（或可比的异步框架）经验，在 MySQL 或 Postgres 上有过硬的 SQL 能力，以及 AWS 的实用知识 熟悉服务端渲染前端（Jinja2、Tailwind、Alpine.js 或 htmx）；我们不寻找 React 专家 精通 Linux 开发，并习惯从 Windows 工作站通过远程桌面进入安全环境工作 在日常工作中每天使用 AI 编码工具，并对其帮助之处和失败之处有清晰的观点 至少有一个交付给真实用户的 LLM 驱动功能，包括你如何评估它以及出了什么问题 对提示词和上下文工程、检索、嵌入和结构化输出有实用知识 具备医疗数据交换标准和基于文件的集成的实用知识：NCPDP D.0、X12 270/271、SFTP、定宽和分隔格式，以及对账 直接沟通者，能坦率地说明权衡，并在不清楚时主动询问 加分项 医疗或其他受监管数据经验（HIPAA、PHI、SOC 2） 药房福利、理赔或临床项目

⚑ 岗位信息收录自公开招聘页（Himalayas），以雇主官网为准