Applied AI Architect

AI is the biggest shift in how organisations build and work in a generation. The organisations who win will be the ones who can move beyond a working demo and get genuinely production-grade, multi-agent AI systems live, reliable and maintained at scale. That is the hardest part of the journey, and it's where this role sits.  As an Applied AI Architect working in Version 1's Forward Deployed capability, you bring deep, hands-on expertise in designing, building, deploying and maintaining agentic systems in production, and you bring it to customers as a trusted adviser, not from behind a desk. You've done this for real — not just prototyped an agent, but taken a multi-agent architecture through the full lifecycle: design, orchestration, evaluation, deployment, and the unglamorous work of keeping it reliable once real users depend on it. You sit alongside customer architects and senior engineers, on their most technically demanding challenges, combining deep technical credibility with genuine consulting skill to shape decisions rather than simply hand over a design.  This is a deeply technical role that lives in the room with the customer. You'll set the technical standard for Version 1's most complex agentic engagements, and you'll do it through influence and trust as much as through the strength of your own architecture.  What you'll do  • Advise the customer's own AI architects. Act as the trusted technical adviser to the customer's AI architect and senior engineers, shaping their design and build decisions on agentic systems rather than owning delivery yourself.  • Consult across the engagement. Run discovery, workshops and architecture reviews with customer stakeholders, translating technical trade-offs into language that lands with both engineers and the business sponsors funding the work.  • Architect multi-agent systems, hands-on. Work directly with customer engineering teams to co-design multi-agent architectures — orchestration, tool use, memory, routing and guardrails — for real production workloads.  • Build reference implementations and accelerators. Create working prototypes, reference architectures, evaluation suites and accelerator code that customer teams take on and run themselves — proof that survives contact with production, not slideware.  • Solve the hard production problems. Bring deep experience of what breaks agentic systems at scale — hallucination under load, tool-call failures, runaway agent loops, latency and cost — and help customer teams design around it.  • Set the technical bar for the engagement. Own the technical architecture and standards for Version 1's most complex agentic engagements, reviewing and pressure-testing customer designs as a peer to their best architects.  • Own evaluation and observability. Establish how a customer will know their agentic system is actually working — eval suites, tracing, guardrails and monitoring — before it goes anywhere near production.  • Mentor on deep technical practice. Coach Applied AI Consultants and Associates on advanced AI engineering, sharing hands-on expertise that goes well beyond the basics.  • Build Version 1's technical IP. Contribute reusable architectures, patterns and accelerators to Version 1's shared toolkit, keeping it current with the frontier of agentic AI engineering.  • Stay genuinely current. Track the latest agent frameworks, orchestration patterns and models closely enough to have a real, tested opinion on what's worth using and what isn't.  What good looks like   • Proven agentic systems experience. You've personally designed, built, deployed and maintained multi-agent systems in production — not a hackathon demo, a system real users depended on.  • A genuine consultant, not just an engineer. You build trust quickly with c