AI Infrastructure Engineer

The impact you will have: This is an opportunity to join Elliptic's AI Platform team at its inception to help build the foundational infrastructure that will power how Elliptic's products think, reason, and act. You will be one of the first engineers working on a centralised AI platform whose purpose is to make AI development faster, safer, and more coherent across the business. That means building the plumbing: the pipelines, the tooling, the evaluation harnesses, the observability layers, and the integration patterns that domain teams will rely on to ship with confidence. You don't need to have done all of this before. What matters is that you are genuinely energised by AI, that you think carefully about how systems fit together, and that you take real pride in building things that others can build on top of. This is a role where curiosity and learning velocity matter as much as prior experience, and where the work you do in the first year will have a lasting shape on how AI, both internally and customer-facing, is engineered at Elliptic. What you will do: - Build and maintain core components of Elliptic's AI platform: model serving infrastructure, prompt pipelines, evaluation harnesses, and integration patterns that allow domain teams to use AI reliably and at scale - Support the development of agentic workflows, including tooling, orchestration scaffolding, and reliability mechanisms, as Elliptic moves toward more autonomous AI capabilities in its products - Instrument AI systems for observability: tracing model calls, tracking token costs, surfacing latency and quality signals, and contributing to the dashboards and alerting that keep production AI systems healthy - Contribute to the tooling and frameworks that govern how prompts are written, versioned, and tested across the organisation, helping to raise the baseline quality of AI interactions across teams - Work closely with engineers in domain teams, such as our Real-time Risk, Investigations, and Data Fabric teams, to understand their AI integration needs and help them build on platform foundations rather than around them - Keep pace with a rapidly evolving AI landscape: new model capabilities, emerging orchestration patterns, and evaluation techniques. Bring relevant developments to the team's attention and help assess what matters for Elliptic's context You will be a great fit here if you: - Are deeply curious about AI. This goes beyond simple tool use and extends to a passion for the field. You follow new model releases, read about emerging architectures, and find yourself thinking about AI applications unprompted - Take pride in building infrastructure that other engineers love to work with. You care about documentation, reliability, and the experience of your internal customers. - Are comfortable with ambiguity and learning in public. You don't need a perfect brief to get started, and you ask good questions when you're unsure rather than guessing quietly - Think holistically about how complex systems interact. You might not yet have built a production AI platform, but you reason well about dependencies, failure modes, and what makes something extensible versus brittle - Are a collaborative and direct communicator. You share what you know, flag what you don't, and make the engineers around you more effective Our ideal candidate has: - Some hands-on experience building with LLMs or ML systems, whether in production, in side projects, or in an academic context. What matters is that you have gone deep enough to understand how these systems actually behave - Familiarity with AI protocols (MCP, A2A, ACP) with a passion to stay current with emerging trends in the industry - Solid software engineering fundamentals: you write clean, testable code, you think about maintainability, and you understand what it means to build something that will be operated in production - A deep understanding of the context window and an appreciation for its importance in extracting maximum v