资深应用AI工程师
Staff Applied AI Engineer
**关于 Rocket Money🔮**
Rocket Money 的使命是显著提升数百万人的财务状况。Rocket Money 为会员提供对其财务的独特理解以及一系列有价值的的服务，帮助他们节省时间和金钱——最终让他们在财务旅程中占据优势。
**关于团队 🤝**
Rocket Money 的使命是显著改善数百万人的财务生活。我们帮助客户了解他们的钱，减少浪费，及时处理账单和订阅，并更有信心地采取行动。
Agents Group 正在构建代理系统，使 Rocket Money 从有用的产品界面发展为可以安全可靠地为客户完成工作的助手。该团队负责启用代理功能的系统以及让代理通过各种联系方式（电话、电子邮件、网络机器人等）与客户和系统交互的方法。
这项工作位于 LLM 驱动的推理、确定性执行系统、工具编排、产品限制、可观测性、评估、信任与安全以及消费级产品设计的交汇点。标准不是“代理能否回应？”，而是“代理能否正确、安全且可验证地完成工作？”
**关于职位**
作为应用 AI 工程师，你将帮助构建、演进和运营 Capabilities Engine：核心平台，使代理行为在 Rocket Money 的 AI 驱动体验中变得可靠、可重用、可观察并适合生产环境。
你将创建能力团队所需的原始工具，将模糊的产品需求转化为可发布的技术方案、原型和持久的系统。你还将与电话、电子邮件、聊天、网页登录、ODC 和自主工作流等联系方法的所有者合作，使每种方法更具代理性和更好地与 Capabilities Engine 集成。这意味着帮助代理规划、发现和调用能力，从模糊中恢复，处理现实中的边缘情况，并安全正确地完成面向用户的工作。
这个职位适合对 AI 作为执行层而非仅对话层充满热情并有经验的工程师。你将构建系统，让代理使用工具、观察结果、从失败中恢复，并获得足够的信任代表用户行事。
你将在产品和平台边界之间工作，跨越架构、实现、调试、评估和跨职能协作。
查看英文原文
**ABOUT ROCKET MONEY🔮**
Rocket Money’s mission is to meaningfully improve the financial prosperity of millions of people. Rocket Money offers members a unique understanding of their finances and a suite of valuable services that save them time and money – ultimately giving them a leg up on their financial journey.
**ABOUT THE TEAM 🤝**
The Agents Group is building the agentic systems that let Rocket Money move from helpful product surfaces to assistants that can safely and reliably finish work for customers. The group owns the system responsible for enabling agentic capabilities and the methods that let agents interact with customers and systems across various contact methods (phone, email, web bots, etc) and autonomous workflows.
This work sits at the intersection of LLM-powered reasoning, deterministic execution systems, tool orchestration, product constraints, observability, evaluation, trust and safety, and consumer-grade product craft. The bar is not “can the agent respond?” The bar is “can the agent correctly, safely, and verifiably complete the work?”
**ABOUT THE ROLE**
As an Applied AI Engineer, you will help build, evolve, and operate the Capabilities Engine: the core platform that makes agentic behavior reliable, reusable, observable, and production-ready across Rocket Money’s AI-powered experiences.
You will create the primitives that capability teams need to turn ambiguous product needs into shippable technical plans, prototypes, and durable systems. You will also partner with contact-method owners across phone, email, chat, web login, ODC, and autonomous workflows to make each method more agentic and better integrated with the Capabilities Engine. That means helping agents plan, discover and invoke capabilities, recover from ambiguity, handle real-world edge cases, and complete user-facing work safely and correctly.
This role is for an engineer who is excited by—and experienced in—AI as an execution layer, not just a conversation layer. You will build systems that let agents use tools, observe outcomes, recover from failure, and earn enough trust to act on behalf of users.
You will work across product and platform boundaries, moving between architecture, implementation, debugging, evaluation, and cross-functional collaboration. Success means making the Capabilities Engine easier to extend, operate, and integrate with; improving task completion and recovery across methods; and ensuring new capabilities ship with clear contracts, tests, traces, ownership, and production-readiness expectations.
**IN THIS ROLE, YOU'LL:**
- Build, evolve, and operate the Capabilities Engine as a core platform for agentic product behavior.
- Design abstractions that make capabilities easy for agents and methods to discover, invoke, compose, observe, and improve.
- Partner with method owners across all contact method subagents as well as supporting agents to make each method more agentic and better integrated with the Capabilities Engine.
- Identify repeated patterns across methods and turn them into reusable capabilities, interfaces, evaluations, and operating practices.
- Provide guidance on how to monitor and optimize reliability, latency, cost, observability, and safety of capability execution in production.
- Develop practical evaluation loops for agentic behavior, including offline tests, production metrics, traces, regression suites, and qualitative review.
- Debug complex failures across model behavior, tool execution, orchestration, product constraints, and downstream systems.
- Translate ambiguous product needs into shippable technical plans, prototypes, and durable systems.
- Create the primitives capability teams need to successfully produce shippable technical plans, prototypes, and durable systems.
- Collaborate with product, design, data, platform, and operations partners to ensure capabilities are useful, understandable, and safe for real users.
- Mentor and develop engineers across the Agents Group, acting as a technical multiplier by sharing expertise, fostering a culture where others feel empowered to propose ideas and take on high-impact work, and helping shape the group's overall technical direction and standards.
- Document system behavior, capability contracts, tradeoffs, and operational expectations so other teams can build on the platform confidently.
- Drop into high-priority method work when needed to accelerate agentic behavior, unblock launches, or stabilize production issues.
- Stay curious about the frontier of applied AI engineering and help the team adopt new tools, patterns, and workflows where they make the product more dependable.
**YOU MAY BE A GOOD FIT IF YOU:**
- Have experience shipping production software with strong engineering fundamentals.
- Are excited by AI as an execution layer. You understand that durable customer value comes from orchestration, tools, workflows, evaluation, approvals, observability, and recovery.
- Have hands-on experience or strong practical familiarity with LLM applications, agentic systems, tool calling, retrieval, workflow orchestration, or applied AI products.
- Have strong product judgment. You can reason from customer need to product experience to technical architecture, and you know the difference between a clever demo and a dependable customer-facing system.
- Are comfortable working across ambiguity. You can take a broad problem, identify the crux, sequence the work, make tradeoffs, and keep moving without waiting for perfect clarity.
- Like building platform abstractions and shared services that help multiple teams or product surfaces move faster.
- Hold a high bar for quality, reliability, and trust. You know that systems acting on behalf of users need clear contracts, observability, evaluation, and recovery paths.
- Are energized by debugging messy production behavior across model outputs, distributed systems, product logic, user context, and downstream integrations.
- Communicate clearly with engineering, product, design, data, operations, and platform partners.
- Are motivated to actively mentor and grow others.
- Are a tenacious learner who stays engaged with the frontier of AI engineering without chasing every shiny object.
- Are motivated by measurable customer and operational impact, not just interesting technical demos.
**STRONG CANDIDATES MAY HAVE:**
- Experience building production LLM applications, agentic systems, workflow engines, tool-calling systems, or AI-powered product surfaces.
- Experience designing capability platforms, shared services, orchestration systems, or reusable product infrastructure.
- Experience with TypeScript, Node, GraphQL, Postgres, LLM applications, agentic systems, and workflow orchestration.
- Experience creating evaluation and observability systems for probabilistic or agentic behavior.
- Experience improving reliability, alerting, tracing, incident response, operational quality, or other systems that make software safer and more dependable over time.
- Experience partnering deeply with Product, Design, Data, Operations, Legal, Support, or other cross-functional teams in a sensitive customer domain.
- Firsthand experience building products where customer trust, permissions, explainability, and recovery paths are central to the experience.
- A visible habit of turning repeated one-off work into reusable systems, patterns, and documentation.
**WE OFFER 💫**
- Health, Dental & Vision Plans
- Competitive Pay
- 401k Matching
- Unlimited PTO
- Lunch daily _(in-office only)_
- Snacks & Coffee _(in-office only)_
- Commuter benefits _(in-office only)_
Additional information: Salary range of $200,000 - $270,000/year + bonus + benefits. Base pay offered may vary depending on job-related knowledge, skills, and experience.
_This job description is an outline of the primary responsibilities of this position and may be modified at the discretion of the company at any time. Decisions related to employment are not based on race, color, religion, national origin, sex, physical or mental disability, sexual orientation, gender identity or expression, age, military or veteran status or any other characteristic protected by state or federal law. The company provides reasonable accommodations to qualified individuals with disabilities in accordance with applicable state and federal laws._
_The information regarding compensation and other benefits included in this paragraph is the company’s current, good faith estimate at the time of posting. [Compensation and benefits are subject to modification from time to time as the Company, in its sole and exclusive discretion, deems appropriate.] The Company may determine during its future reviews of the proposed compensation and benefits provided for this position, that the compensation and benefits for such position should be reduced. In no event will the Company reduce the compensation for the position to a level below the applicable jurisdictional minimum wage rate for the position. Los Angeles County and San Francisco Candidates only: qualified applicants with arrest or conviction records will be considered for employment per the Fair Chance Ordinance and the Fair Chance Initiative for Hiring._