AI Solution Architect

Lorien; Greater London; Hybrid; GBP 110,000 - 150,000; Full time

Job summary

Lorien is seeking an AI Solution Architect to design secure, scalable, production‑ready AI and agentic solutions in London with hybrid work. You will bridge business requirements and engineering design, deciding what stays deterministic, what is probabilistic, and what needs a human in the loop. You will guide cross‑team decisions and ensure governance alignment.

Qualifications

• Hands-on architect with strong cloud, integration and distributed-systems experience.
• Practical knowledge of LLMs, agents, retrieval and evaluation.
• Problem-solver with experience across different businesses.

Responsibilities

• Design end-to-end architectures for agent and workflow automation.
• Select models, orchestration patterns, data sources and integration approaches.
• Define boundaries between deterministic software, probabilistic AI and human decision-making.
• Address identity, permissions, data protection, resilience, observability and cost from day one.
• Design evaluation, fallback, escalation and human-in-the-loop mechanisms.
• Produce architecture decisions, patterns and reusable reference designs.
• Review implementations and guide engineering teams through major technical decisions.
• Ensure solutions align with enterprise architecture and governance requirements.

Skills

Cloud architecture

Distributed systems

LLMs / agentic AI

System integration

Security & governance

Job description

AI Solution Architect, Agentic & LLM Solutions

Location: Hybrid -

London, 3 days a week

IR35 Status:

Inside

Summary:

We're looking for a hands‑on solution architect to design secure, scalable, and genuinely operable AI and agentic solutions, the kind that survive contact with production, not just a demo.

You’ll sit at the intersection of business requirements, product intent and detailed engineering design, and you’ll be the one drawing the line between what should be deterministic software, what should be probabilistic AI, and what still needs a human in the loop.

What you’ll be doing:

• Designing end-to-end architectures for agent and workflow automation
• Selecting appropriate models, orchestration patterns, data sources and integration approaches
• Defining the boundaries between deterministic software, probabilistic AI and human decision‑making
• Addressing identity, permissions, data protection, resilience, observability and cost from day one not bolted on later
• Designing evaluation, fallback, escalation and human-in-the-loop mechanisms
• Producing architecture decisions, patterns and reusable reference designs
• Reviewing implementations and guiding engineering teams through significant technical decisions
• Making sure solutions align with wider enterprise architecture and governance requirements

What we’re looking for:

• A hands‑on architect with strong cloud, integration and distributed‑systems experience
• Practical, real knowledge of LLMs, agents, retrieval and evaluation — not just conceptual familiarity
• A problem‑solver mindset with a clear memory of what architectural approaches have worked (and what hasn’t) across different businesses
• Comfort working across multiple business areas rather than one narrow domain
• MBB consulting background (McKinsey, Bain, BCG) is a strong plus

How this works:

• Open to contract or permanent hire
• UK‑based candidates: ideally on‑site 3 days a week in London
• Candidates based outside the UK: on‑site 3 days every fortnight, with flights and accommodation covered for the right candidate with consulting‑firm pedigree
• All contract engagements sit inside the scope of IR35