Principal AI Architect

About Invoka Consulting

Invoka Consulting partners with clients across all industries and around the world to design, implement, and optimise innovative SaaS and AI solutions that deliver measurable business impact. We are an Elite Smartsheet partner, OpenAI and an Anthropic partner.

Our AI work is already in production, not on a roadmap. We have automated a regulatory compliance process for a national sporting body that now assesses every case in hours rather than weeks, we run daily AI workloads inside a global pharmaceutical business, and we have built our own Model Context Protocol server for Smartsheet with roughly ten times the tool surface of the public one.

We are now expanding our AI practice.

Overview of the Role

We are looking for a Forward-Deployed Principal AI Architect to serve as the technical anchor of Invoka's AI practice. This is a hands-on role. You will spend your time designing and building production AI systems, embedded with enterprise clients, and turning what we learn on each engagement into reusable capabilities.

We do not believe enterprise AI works when it is specified from the outside and handed over. Our delivery model puts engineers inside the client's business to understand how the work genuinely runs across departmental seams, and then to build against that reality. You will be the person who proves that model works, and the person who sets the standard everyone hired after you is measured against.

Alongside the delivery work, you will own our AI architecture standards, our evaluation and reliability practices, and our accelerator framework: the reusable components, prompt libraries and integration patterns that turn one-off projects into repeatable delivery. You will lead technical pre-sales on our most complex opportunities, and you will set the technical bar for the AI roles we hire in the UK and other geographies.

This is a role for someone who wants ownership. You are joining a practice early enough to shape what it becomes, with real enterprise clients to build for from day one.

Key Responsibilities

Architecture and Technical Standards

• Design and sign off the solution architecture on AI engagements, working within Invoka's Responsible AI framework.
• Establish and own our evaluation, testing and reliability standards for AI systems, so that performance is measured rather than assumed.
• Set the technical standards, patterns and documentation practices the AI practice delivers against.
• Stay at the frontier across Anthropic, OpenAI and Microsoft, and translate capability changes into delivery capability and commercial offerings.

Building and Embedded Delivery

• Build production AI systems hands-on: agents, tool use, retrieval pipelines, MCP integrations and enterprise workflow connectors.
• Embed with client teams to map how work actually runs across functions, then design and deliver against it.
• Lead the delivery of a flagship cross-functional workflow end-to-end, as the reference case for the practice.
• Act as the escalation point for complex technical issues on live engagements.

Capability and Intellectual Property

• Own and productise Invoka's AI delivery IP: the accelerator framework, reusable components, prompt libraries and evaluation suites.
• Take our in-house MCP servers from internal tools to supported, documented, sellable asset, and own its roadmap.
• Develop our temporal data and change analysis capability, which we deliver as a service inside client engagements, where it intersects with the AI practice.

Pre-Sales and Client Engagement

• Lead technical pre-sales on complex and high-value opportunities: discovery sessions, solution shaping, technical scope and feasibility.
• Own the technical narrative in front of Director, VP and C-suite stakeholders, including sceptical security and legal functions.
• Contribute to proposals, RFP responses and Statements of Work, ensuring what we sell is what we can deliver.

Practice and Team Development

• Set and hold the technical bar for AI delivery hires: define the interview process, assess candidates, and mentor the engineers who join.
• Write the technical onboarding for our Poland delivery hub, launching next year.
• Contribute to case studies, technical writing and partner assets that build Invoka's reputation in the market.

Required Skills and Experience

• Production AI systems you have actually shipped. You have designed, built and deployed LLM-based systems that run in production with real users, uptime expectations and sensitive data. We will want to talk about who uses them, how often, and what broke.
• Evaluation and reliability engineering. You can explain how you measure whether a non-deterministic system is working, and how you would know six months later that it had degraded. Evaluation design, regression suites, golden datasets, failure taxonomies, guardrails and human review loops are familiar territory.
• Model Context Protocol at authoring depth. You have written or substantially extended MCP servers, or equivalent agentic tool frameworks. We already run MCP servers in production, so we need someone who can extend, productise and maintain them.
• Enterprise integration, security and data governance. Deep experience connecting AI to enterprise platforms, and fluency in what a CISO and a legal team will actually ask about data residency, PII handling, authentication and boundary crossing.
• Python or TypeScript to a professional standard. Readable, tested, documented code.
• Reusable assets, not just projects. You have built IP inside a consulting or product environment and understand that unversioned, undocumented work is not an asset.
• Enterprise client presence. Credible and comfortable with Director, VP and C-suite stakeholders. You can explain an architecture to a sceptical CISO and a cost model to a CFO in the same meeting. Clear writing matters here as much as clear speaking.

Nice to Have

• Smartsheet development experience: API, Bridge, Data Shuttle. This is the intersection where much of our AI work sits.
• Temporal or slowly changing dimension data modelling.
• Multi-vendor fluency across Anthropic plus OpenAI, Microsoft Copilot or Azure OpenAI. Our advisory work is deliberately tool-neutral.
• Experience in regulated sectors: pharmaceuticals and life sciences, financial services, or public sector.
• Technical leadership of offshore or nearshore engineering teams.
• Published technical writing, conference speaking, or open source contributions.

What the First Year Looks Like

• First 90 days. Embedded as technical lead on a live enterprise engagement. Architecture standards and the evaluation framework documented and in use. Named technical lead on live proposals.
• First six months. Your first accelerator shipped and reused on a second engagement, with the delivery time saving measured. A flagship cross-functional workflow delivered end-to-end. Owning the technical interview process for the AI team.
• First twelve months. An accelerator library that measurably reduces delivery effort on repeat work. Architecture sign-off embedded across the practice. The technical onboarding written for our offshore hub. A track record of enterprise work won with you as the named technical lead.

Perks and Benefits

• Equity participation.
• Performance-related bonus.
• Employer-paid private medical and Dental insurance.
• Monthly employer pension contributions.
• 25 days paid holiday plus 8 UK bank holidays plus your birthday off.

Initial screening for this role is being managed by our third-party recruitment partner, AES. Successful candidates will be contacted by AES, and all interviews are conducted by Invoka's leadership team.