Staff Machine Learning Engineer

Staff Machine Learning Engineer — AI-native productivity, stealthFully remote, UK-based candidates preferred.TL;DRFounding-team Staff MLE at a well-funded stealth AI companyProduct: AI-native productivity — starting with email, expanding into notes, tasks, calendarOwn the full model lifecycle: data, training, evaluation, inference, deploymentUS$100M initial funding, internally backed, no VC pressureCash + meaningful founding-team equityFully remoteThe playEmail, calendar, notes, tasks. The tools 5 billion people run their lives on. None of them are AI-native. Every attempt so far has been a bolt-on — a copilot button, a summary at the top of the thread. This company is building the layer underneath: proactive, context-aware, capable of running long workflows, completing real tasks, and asking before it acts. First product is AI-native email. The goal: cut four hours a day in the inbox down to thirty minutes. Email first. Productivity suite next.The role, first 12 monthsOwn the execution layer of the company's intelligence — turning research and model capabilities into reliable, scalable production systems. Build and evolve fine-tuning pipelines for large models. Design evaluation systems that measure real-world capability, robustness, and safety — not benchmark vanity. Architect high-performance inference infrastructure: latency, GPU utilisation, memory, cost. Build data pipelines for high-quality real-world and synthetic training data. Bridge research and application engineering so model improvements actually reach users.The barRead this before you DM.Production ML systems you've built and shipped — not prototypes, not research demos, real products with real usersDeep understanding of large-model training, fine-tuning, evaluation, and inferenceExperience running GPU-based ML workloads at meaningful scaleStrong software engineering fundamentals — you write production-grade code and you care about correctnessComfortable reasoning about failure modes, model degradation, and what happens when things go wrong in the wildIndependent judgment — you can navigate ambiguity and make pragmatic trade-offs without being hand-heldWho this isn't forML engineers who've only ever worked in notebooks. Researchers chasing publications. Anyone who needs a stable, well-defined system before they can contribute. This is a hands-on, high-ownership role in a pre-launch team moving fast.If the bar above doesn't quite match where you are today — no worries. Save your energy for the role that does.The restEverything else — who they are, who's behind it, comp and equity detail — is a call.DM me if this is you, or if you know the person it should be.