Forward Deployed Engineer

The role We're building a customer-facing AI engineering function inside Customer Success to turn AI conversations into signed pilots. Forward Deployed Engineers sit alongside prospective and existing customers — from the first exploratory meeting through to a live pilot — and are the technical reason a customer says yes. The work spans two areas, and you will do both. Applied AI engineering: prototypes and custom demos, reference deployments and proofs-of-concept, GPU sizing and architecture, inference benchmarking and model fit. Business value engineering: TCO modelling and commercial structuring, tender and pre-sales support, CXO and strategic engagements, thought leadership and GTM. You will build, and you will build quickly — but what you build exists to win and shape a deal, not to run in production for three years. Production hardening and steady-state run sit with the AI infrastructure and production team. Roughly half this job is technical judgement and half is communication and commercial reasoning. What you'll do • Scope the opportunity. Work with Customer Success and Sales to understand a prospective customer's problem, workflow and constraints — then decide what to build that will move the deal forward. • Build the prototype or custom demo that wins the deal. Stand up working prototypes, custom demos and reference deployments on RE:AI — agentic workflows, retrieval, tool use, guardrails — tuned to the customer's actual use case rather than a generic demo. Demonstrate them live wherever possible, not as a slide deck. • Size the platform and prove the fit. GPU sizing and architecture, inference benchmarking, and model selection against the customer's latency, throughput and cost targets — and the ability to show your working when their architects ask how you arrived at the number. • Support the pilot. Stay technically engaged as a proof-of-concept converts to a pilot: integration guidance, tuning, and being the escalation point the customer trusts — handing over to the AI infrastructure and production team for build and run once the decision is made. • Build a reusable asset library. Patterns that show up across three customers should become a reusable demo, template or accelerator, not three bespoke builds. You'll be expected to spot that and push components back into shared pre-sale tooling. • Be the technical face of RE:AI. Represent the platform credibly to customer engineers, architects and CXOs — explaining what you built, why, and what tradeoffs you made, in their language. • Support tenders and formal pre-sales. Technical content for RFPs, government tenders and compliance responses — solution write-ups, architecture sections and clarification responses, working with bid and commercial teams to deadline. • Model the commercial case. TCO modelling, token economics and commercial structuring — quantify what a solution costs to run on RE:AI and what it is worth against a public API or on-premise alternative, in terms a CFO would accept. • Contribute to thought leadership and GTM. Reference architectures, customer-facing write-ups, webinars and conference material that make the next deal easier to open — plus internal enablement so Sales can carry the story without you in the room. What we're looking for • Genuinely customer-facing. You can run a technical conversation with a customer's engineers directly, hold your position under pushback, and read a room well enough to know when to go deeper and when to simplify. This is the single most important requirement. • AI-native. You have personally built with LLMs and agents — tool-calling, retrieval, orchestration, prompting, evaluation — and you keep up with the field because you're interested in it. Certifications alone won't cover this. • Excellent communicator, spoken and written. You can present to a mixed technical and executive audience, and write a clear proposal, tender response or follow-up that stands on its own. Your writing will be seen by customers. • Commercial and value judgement. You think about business value, not just the software. You can build or reason about a TCO model, explain where cost sits and what drives it, and articulate why a customer should choose this — and you can say when the right answer is a smaller solution. • Comfortable with ambiguity and travel. Every customer and every pilot is different. You'll define your own scope more often than you're handed one, and you're fine being on a plane or on-site when the deal needs it. • Sound engineering fundamentals. Strong enough in at least one backend stack to build a credible prototype yourself, comfortable with APIs and cloud basics, and able to reason clearly about how a model behaves in production — latency, cost, failure modes. You don't need to have trained a model or run a production platform. Nice to have • Product management or product owner experience — the ability to frame a problem in terms of user value and business outcome. • Full-stack capability, enough to make a prototype usable end to end without waiting for anyone. • Experience responding to enterprise RFPs or public sector tenders, or building TCO and commercial models. • Familiarity with GPU infrastructure, inference serving, or benchmarking model performance and cost. • Prior forward-deployed, solutions-engineering, or startup generalist engineering experience. • Experience in regulated industries (financial services, healthcare, government) — understanding what "production-ready" means when compliance and data residency are non-negotiable. • Experience with LLM gateways/routing layers, guardrail/policy systems, or building on top of an internal model-serving platform. • A track record of turning one customer's build into a reusable pattern for the next three.