AI Operations Engineer

About us OpenZeppelin is the security standard onchain finance is built on. Founded in 2015, our mission is to accelerate the world’s transition to an open financial system, built on open standards and secured by rigorous research. Our open-source Contract Libraries have facilitated over $35 trillion in onchain value and are used by 10 of the top 10 tokenized money market funds and 9 of the top 10 stablecoins by market cap. We combine AI-native security tooling with deep research and a decade of audit expertise to support leading institutions and crypto-native teams shaping the next generation of digital assets like DTCC, Fidelity, Coinbase, Uniswap, Aave, the Ethereum Foundation, and many more across the full secure development lifecycle. Please note: Always refer to OpenZeppelin’s official job page for the most accurate information about our open roles, as we have seen multiple third party job sites posting inaccurate information. The AI Operations Team The AI Operations team owns the agentic layer the rest of the company runs on. We build the workflows, agents, and scheduled tasks that automate business processes end to end, and we build and maintain the proprietary framework those components ship in. Every business unit at OpenZeppelin is a customer of this team. What you’ll be doing The role has two primary work streams. AI workflow design, implementation, and maintenance. You understand the fundamental AI building blocks that we use in the company, and you work with interdisciplinary business units to help them identify flows in their processes that can be optimized and automated with AI. Internal AI automation framework product feedback and contribution. Based on the experience that you obtain as a consumer of the internal AI automation framework, and on your study of the ecosystem, you provide product feedback (feature requests, bug reports, etc.), contribute to the source code of the framework, and take part in peer review and development. Work together with different business units in the company to understand where and how AI can be applied in their daily workflows to improve their overall productivity. Design, implement, test, and maintain AI automation workflows and scheduled tasks that encapsulate and automate business processes that span the company. Monitor and measure automation executions, and iterate on the components that underperform. Provide product feedback for fixes and new features in the internal AI automation framework, and peer review and actively maintain its source code as a developer on the AI Operations team. Run training sessions and write internal guides on applied AI topics such as skill development, agent design, and prompt/context engineering. You have Hands-on experience building applied AI systems, and fluency with both the concepts involved (prompt engineering, context engineering, prompt caching, tool use, progressive disclosure) and the fundamental building blocks: skills, remote and local MCP servers, agents, and agent harnesses. Practical experience with LLM provider APIs and gateways, such as the the OpenAI Chat Completions and Responses APIs, the Anthropic Messages API, or OpenRouter. Ability to read, write, and structure Markdown-based systems (skills, agent definitions, workflow specs, and knowledge bases), alongside TypeScript development experience and solid Git and GitHub fluency: branches, pull requests, reviews, and the discipline to keep shared repositories clean. A security-first mindset: you think about what an autonomous agent should not be allowed to do before you ship it. Strong ability to communicate deeply technical topics, both written and spoken, and patience working with non-technical colleagues. You will be their bridge into the agentic workflow. Nice to have Familiarity with AI software development techniques like spec-driven development or RPI (Research, Plan, Implement), with the principal AI development harnesses like Claude Code, Codex, or OpenCode, and with SDD tooling like Superpowers or Spec Kit. Experience running evaluations of LLM systems (evals, benchmarks, regression testing of prompts). Ability to work autonomously and as a collaborator on the team. Logistics: Our interview process takes place on Google Meet or Zoom and tends to consist of the following stages: Recruiter Call (30 minutes) Hiring Manager Call (60 minutes) Team Interview (30 minutes) Leadership Interview (30 minutes) Paid work trial (10-20 hours of paid work) Reference checks Benefits Meet your teammates at company gatherings around the world 😎 Enjoy the flexibility of fully remote work 🌎 Take the time you need with flexible time off 🏝 Grow your family with 8 weeks of paid leave for primary caregivers, 4 weeks for secondary caregivers, and a one-time $3,600 baby bonus 💙 Build your ideal home office with up to $500 in equipment support 🪑 Stay covered with medical insurance 🏥 Keep growing with learning and development opportunities 🧠 Get a monthly stipend for your preferred co-working space 💻 At OpenZeppelin , we are an equal opportunity employer and we value different perspectives. We are committed to building a diverse workforce. This includes but is not limited to gender, race, sexual orientation, religion, national origin and other characteristics that make each one of us unique. In this uniqueness, we find the most value. Come join us! Use of AI as part of the recruiting process As part of OpenZeppelin’s recruitment process, we may use automated tools, including artificial intelligence, to assist in reviewing applications and assessing candidate qualifications. These tools are used to support our People team by identifying relevant skills and experience, and are not used to make decisions solely by automated means. All hiring decisions involve human review. Any personal data provided as part of your application will be processed in accordance with OpenZeppelin’s Data Privacy Notice. If you have questions about this recruitment process or would like to request human review of your application, please contact us at [email protected].