高级工程师 - 智能体
查看雇主原标题
Senior Engineer - AgentsBuildkite · ANZ Region
职位信息来自雇主公开的招聘页面。申请前请务必在雇主官网核实详情。
为什么值得关注？
发现指数 49/100，仅依据与该职位一起存储的证据计算。
- 新的雇主官方职位
- 稀有职位匹配
分数构成
- 时效性 （随职位发布时间变化）+18
- 雇主官方来源+15
- 稀有职位+8
- 公司来源健康度+8
该职位未包含：已披露薪资、远程职位、提及签证担保、提及搬迁、未出现在监控的职位板上。
这些理由来自雇主自己的职位描述与我们核实过的来源检查结果。除了已存储的信号之外，我们不做任何推测。
职位描述
机器翻译关于 Buildkite
Buildkite 的 CI 平台深受全球领先工程团队的信任，帮助他们交付每天有超过 10 亿人使用的软件。Canva、Meta、Mistral AI 和 NVIDIA 等公司都依赖 Buildkite 在他们自己的基础设施、我们的基础设施或两者之上运行构建、测试和部署——让他们拥有速度、可靠性和灵活性，从而自信地交付。
岗位职责
此刻，某处正有一位工程师推送了一次变更，并正在看着一次构建运行。完成这项工作的就是 Buildkite Agent——而这正是这个团队所负责的。它是那个开源 agent，世界各地的工程师都在自己的机器上、自己的云中、Buildkite 自己的托管 agent 中运行它，凡是 Buildkite 的 CI 运行的地方，无处不在。
这是一个至关重要的代码库。这个团队处在一个令人兴奋的交汇点上——一边是深度技术性的基础设施工作，另一边是快速发展的 AI agent 世界，而我们自己的 agent 夹在中间。它是其他一切所依赖的那一块，所以细节很重要。依赖你的人，是那些正在交付自己功能的工程师同伴。
你将负责的系统不能被暂停或随意重启。我们需要不断迭代和更新。这就是它的本质，也正是它大部分有趣之处。你会有可以指着它并为之自豪的工作成果，以及一个敏锐、有主见的团队，他们会很乐意和你一起花一个下午攻克一个难题。入职几个月后，我们期望你能完全负责某件重要的事情。
交付
• 把大问题拆成本周就能上生产的小事情，而不是八个月后才上生产的完美事情。
• 端到端负责功能——从想法，到权衡取舍，到上线发布，再到它在生产中出现异常时会发生什么。
• 为一个会持续发生局部故障的世界而设计：谨慎的发布、可回滚的迁移，以及你在事故期间真正想要的可观测性。
负责
• 接手一个问题并为之着迷。需要挖多深就挖多深。带着答案和计划回来。
• 轮值 Sweep——每八周中有两周，通过处理进来的事务、与 Support 和 Solutions Engineers 合作，并快速给人答复，来保护团队的项目时间。这是这份工作中那种快速见效、让人高多巴胺分泌的一端。
• 把想法带到路线图中，而不只是执行路线图。如果它比已经计划好的事情更有价值，我们就会把它换进来。
• 处理事故，并在问题的余烬变成麻烦的大火之前将其扑灭。
沟通
• 解释你在做什么、你发现了什么，以及你担心什么——无论是书面还是口头，都要清晰，对工程师和非工程师都一样。
• 结对、评审，并给出让代码和人都变得更好的反馈。
• 像使用电动工具一样使用 AI 工具，而不是自动驾驶：把 agent 扔向问题，然后验证返回的结果。
如果你是……这就是你
• 你更愿意做给我们看，而不是说给我们听。你能指着你构建过的东西、你改变过的东西，以及你负责过的东西——并用第一人称谈论它们。
• 你小步且频繁地交付。面对一个八个月的问题时，你的本能是找到快速原型并把它呈现在人们面前。
• 你会变得着迷。当某样东西坏了、慢了或丑了，你无法置之不理。你会去查明原因。
• 无论如何你都会去做。你可以坦然地说“我以前没做过这个，这让我很不舒服，但我弄明白了怎么做”。而“那不是我的工作”不在你的词汇表里。
• 你是很强的沟通者。你会在别人没要求的情况下让身边的人了解情况。你可以在一个满是同样有主见的人的房间里自在地表达自己的观点。
• 你背后有真正的生产环境负责经验。不是“我在那个团队里”，而是“我负责，这里是哪里坏了以及我改了什么。”
• 你非常精通 Go 或 Ruby，而且你对两者都不执着。Go 是我们在 agent 和基础设施中的主要语言。Ruby on Rails 是平台的骨干。你至少需要其中一种的真实经验——并且要有能力在工具格局在我们所有人脚下变化时，在不同语言之间自如切换。
加分项：具备以下经验；
• Kubernetes 和 AWS——我们大规模运行云原生基础设施。
• Terraform——我们以代码方式管理基础设施，并重视可复用、可维护的模块。
• PostgreSQL——查询优化、schema 设计，或在高负载下运行关系型数据库。
• React 和 GraphQL——我们的前端和 API 层。
• 你做过开发者工具、CI/CD，或任何你的用户是有强烈主见的工程师的东西。
• 你曾公开写过你追查过的问题——一份事后复盘、一次深入剖析、一次“为什么这么慢”的调查。
申请时——带一个故事来
这段对话会很早出现，所以也许可以为此做好准备：
告诉我们你快速构建并交付过的某样东西。不是你做过的最大的东西，但也许是那个你把问题拆解开来、做出决定，并让某个小而有用的东西比任何人预期都更快进入生产的项目。你跳过了什么？你哪里做错了？接下来发生了什么？
这段对话是我们流程的核心。值得提前想一想。
福利待遇
• 每年最多 6 周可以在世界任何地方工作，此外还有完全远程的工作设置——出色的工作不必只在一个办公桌前完成。
• 20 天带薪年假，外加 10 天病假和照护者假——我们积极鼓励休假，我们希望你能处于最佳状态，而休息时间是其中一部分。
• 主要照护者享有 16 周带薪育儿假（次要照护者 6 周）——成为父母会改变人生，我们致力于在你和你的家人度过这一阶段时提供支持。
• 远程工作津贴，用于家庭办公套件、共享办公或在其他地方工作——我们希望你能拥有在岗位上取得成功所需的一切。
• 保密的 EAP 和心理健康支持——有时生活会很艰难。你应该能够获得所需的资源和支持。
• 支持学习和成长——提供参加会议、课程和辅导的时间与预算，让你有机会成长、提升技艺并建立精通。
• 退休和健康保障（根据你所在地区）——以你的方式为你的未来做准备。
具体福利可能因国家而略有不同。我们会详细说明你所在地区的具体情况。
接下来会发生什么
每一份申请都会得到回复。我们的目标是快速推进招聘流程，并让每个人尽可能没有压力。
如果这听起来像你已经在做的工作方式，那就申请吧。如果你先有问题，请联系我们，我们更愿意先聊一聊。 🌈 平等机会雇主
在 Buildkite，我们重视多样性，并赞赏各种类型的技能、背景和经验。我们致力于营造包容的环境，并在整个招聘过程中提供合理的便利安排。
如果你在申请或面试过程中需要任何便利安排或支持，请通过 accommodations@buildkite.com 联系我们。
以上内容由机器翻译自动生成，可能存在错误；投递前请以雇主原文为准。
查看雇主原文
职位描述
About Buildkite
Buildkite's CI platform is trusted by the world's leading engineering teams, helping ship software used by more than 1 billion people every day. Companies including Canva, Meta, Mistral AI, and NVIDIA rely on Buildkite to run builds, tests, and deployments on their infrastructure, ours, or both — giving them the speed, reliability, and flexibility to ship with confidence.
岗位职责
Somewhere right now, an engineer has pushed a change and is watching a build run. What's doing that work is the Buildkite Agent — and that's what this team looks after. It’s the open-source agent that engineers around the world have running on their own machines, in their own clouds, in Buildkite’s own hosted agents, anywhere and everywhere Buildkite’s CI runs.
It's a critical repo to own. The team sits at an exciting intersection — deeply technical infrastructure work on one side, the fast-moving world of AI agents on the other, and our own agent sandwiched in the middle. It's the piece everything else leans on, so the details count. The people relying on you are your fellow engineers shipping their own features.
The systems you'll look after can't be paused or casually restarted. We need to constantly iterate and update. That's the nature of it, and it's most of what makes it interesting. You'll have work you can point at and be proud of, and a sharp, opinionated team who'll happily spend an afternoon in a hard problem with you. A few months in, we'd expect you to be fully owning something that matters.
Ship
• Break big problems into small things that go to production this week, rather than perfect things that go to production in eight months.
• Own features end-to-end — from the idea, through the trade-offs, to the rollout, to what happens when it misbehaves in production.
• Design for a world where things fail partially and constantly: careful rollouts, migrations you can reverse, and observability you'd actually want during an incident.
Own
• Take a problem and get obsessive about it. Dig as deep as it takes. Come back with an answer and a plan.
• Rotate through Sweep — two weeks in every eight, protecting the team's project time by fielding what comes in, working with Support and Solutions Engineers, and getting people answers fast. It's the quick-hit, high-endorphin end of the job.
• Bring ideas to the roadmap, not just execution against it. If it's more valuable than what's already planned, we'll swap it in.
• Handle incidents and put out problem embers before they become problematic fires.
Communicate
• Explain what you're doing, what you found, and what you're worried about — clearly, in writing and out loud, to engineers and non-engineers alike.
• Pair, review, and give feedback that makes the code and the person better.
• Use AI tooling like a power tool, not autopilot: throw an agent at the problem, then verify what comes back.
This Is You If...
• You'd rather show us than tell us. You can point at things you've built, things you've changed, and things you've owned — and talk about them in the first person.
• You ship small and often. Your instinct when facing an eight-month problem is to find the quick prototype and get it in front of people.
• You get obsessive. When something's broken or slow or ugly, you can't leave it alone. You go find out why.
• You do it anyway. You're comfortable saying "I hadn't done this before and it was uncomfortable, but I worked out how". And "that's not my job" isn't in your vocabulary.
• You're a strong communicator. You keep people around you informed without being asked. You're comfortable having an opinion in a room full of people who also have opinions.
• You have real production ownership behind you. Not "I was on the team" but, "I was responsible, and here's what broke and what I changed."
• You are highly proficient with Go, or Ruby, and you're not precious about either. Go is our primary language across the agent and infrastructure. Ruby on Rails is the backbone of the platform. You'll need genuine experience in at least one — and the comfort to move between languages as the tooling landscape shifts underneath all of us.
Bonus Points for experience with;
• Kubernetes & AWS — we run cloud-native infrastructure at scale.
• Terraform — we manage infrastructure as code and value reusable, maintainable modules.
• PostgreSQL — query optimisation, schema design, or running relational databases under load.
• React & GraphQL — our frontend and API layer.
• You've worked on developer tooling, CI/CD, or anything where your users are engineers with strong opinions.
• You've written publicly about a problem you chased down — a post-mortem, a deep dive, a "why is this slow" investigation.
When You Apply — Come With a Story
This conversation will come up early, so maybe be prepared for it:
Tell us about something you built and shipped quickly. Not the biggest thing you've ever worked on, but maybe the one where you broke a problem down, made a call, and got something small and useful into production faster than anyone expected. What did you skip? What did you get wrong? What happened next?
That conversation is the heart of our process. It's worth thinking about beforehand.
福利待遇
• Work from anywhere in the world for up to 6 weeks a year , on top of a fully remote setup — great work doesn't have to happen from one desk.
• 20 days paid annual leave plus 10 days sick & carer's leave — actively encouraged, we want you to be at your best and time off is a part of that.
• 16 weeks paid parental leave for primary carers (6 for secondary) — becoming a parent is life changing and we're committed to supporting you and your family through it.
• A remote-work allowance for home-office kit, co-working, or working elsewhere — we want you to have everything you need to be set up for success in your role.
• Confidential EAP & mental health support — sometimes life gets hard. You should be able to access the resources and support you need.
• Support for learning and growth — with time and budget for conferences, courses, and coaching so you have the opportunity to grow, improve your craft, and build mastery.
• Retirement and health coverage (based on your region) — build towards your future, your way.
Exact benefits may vary slightly by country. We'll walk through the specifics for your location.
What Happens Next
Every application gets a response. We aim to move quickly through our hiring process and make it as stress free as possible for everyone.
If this sounds like how you already work, apply. If you've got questions first, reach out, we'd rather have the conversation. 🌈 Equal Opportunity Employer
At Buildkite, we value diversity and celebrate all types of skills, backgrounds, and experiences. We’re dedicated to fostering an inclusive environment and providing reasonable accommodations throughout our recruitment process.
If you need any accommodations or support during the application or interview process, please reach out to us at accommodations@buildkite.com.