Technical Expert - Agentic Systems in London

The Esko AI team is a group of specialists focused on leveraging AI to solve valuable and often challenging customer problems that have historically resisted automation via purely rule-based approaches. To succeed in the complex domain of packaging graphics, AI agents must operate reliably according to regulatory requirements, technical specifications, brand guidelines, customer-specific conventions - and invoke expert human judgement when required. This is a complex engineering problem that goes beyond the boundaries of standard agentic scenarios.

We're looking for a Technical Expert to take technical ownership of core agentic capabilities within Esko’s AI Platform, from experimentation and design through production implementation. This is a hands-on technical role, not a management one: you'll set direction and raise the bar through writing code, leading experiments, writing documentation, and mentorship of team members for a platform that spans multiple agentic capabilities (knowledge & reasoning, tool use and orchestration, evaluation, and safety/guardrails) and which is used by many product teams across the business.

You should bring experience from multiple agentic AI projects with a similar order of complexity in terms of the technical challenges you have successfully tackled, together with a ‘sixth sense’ for agentic AI operation to predict the approaches which are likely yield successful results as the platform grows into new domains.

What you will do

• Go deep on the hardest technical problems inside the agentic AI platform - the ones where progress depends on specialist expertise, not just engineering effort: data discovery & consolidation, agent planning and tool-use, retrieval and grounding, validation-before-action, conflict/ambiguity handling, and similar hard sub-problems.
• Personally design, build, and harden the platform's most technically demanding components - the pieces where correctness, edge cases, and failure modes are subtle enough to need deep, hands-on ownership rather than delegation.
• Act as the team's go-to technical expert on agentic AI and LLM systems: the person others bring their hardest "how do we actually make this work" problems to, providing depth on specific technical questions rather than setting overall team direction.
• Diagnose and resolve the platform's toughest technical issues - performance, reliability, correctness, and edge-case failures - where specialist knowledge is the difference between a real fix and a workaround.
• Contribute into architecture discussions, stress-testing designs against hard technical realities (scaling limits, failure modes, subtle correctness issues) that come from real specialist experience.
• Evaluate emerging techniques, frameworks, and tools in agentic AI and LLM tooling, and give the team pragmatic, evidence-based recommendations on what's worth adopting.
• Write technical deep-dives and documentation that make complex mechanisms legible to the rest of the team, so your expertise compounds rather than staying tacit.

What we are looking for

• Strong software engineering experience designing, building, and operating production systems in Python, TypeScript, Java, or a similar language.
• 10+ years of software engineering experience, including substantial time as a senior/staff/principal-level individual contributor architecting and operating production agentic AI systems involving multi-step reasoning, tool use, planning, or orchestration, where alignment and dependability of the agent’s output is a key success factor.
• Experience across multiple distinct agentic AI projects, able to draw on that breadth to guide technical decisions in unfamiliar contexts.
• Deep, hands-on expertise in: LLM agent frameworks and orchestration, tool-calling and validation-before-action patterns, RAG and retrieval architecture, evaluation and observability for AI systems.
• A track record of setting technical direction and shaping outcomes across teams without formal management authority.
• Sound engineering judgement on production concerns at platform scale: reliability, security, multi-tenancy, performance, and observability for services consumed by many internal teams.
• Excellent technical communication - able to write RFCs and architectural documents that align diverse stakeholders, and to move fluently between deeply technical and business audiences.
• Comfort operating with significant ambiguity and autonomy, including defining the right problem as much as solving it.
• Undergraduate, Master's, or PhD in Computer Science, Machine Learning, Data Science, Artificial Intelligence, or a related field (or equivalent experience).

Preferred experience

• Public technical-leadership signals: conference talks, published papers or blog posts, significant open-source contributions, or patents in AI/agentic systems.
• Semantic and symbolic tooling (RDF/OWL, SHACL, Datalog, DMN, production rule engines, policy-as-code) as one of several tools in a broader agentic toolkit.
• Experience in high-accountability, regulated domains (pharmaceutical or food labelling, legal technology, financial services) or the packaging/graphic arts industry.
• Experience scaling an AI or data platform from a handful of internal consumers to many, including the governance and support model that requires.

What success looks like

• The platform's hardest technical problems get solved correctly and durably.
• The team turns to you as the deep technical authority on agentic AI and LLM systems - for the "how" and "why it breaks.".
• The components you build are robust, well-tested, and well-documented, with edge cases and failure modes properly accounted for rather than discovered in production.
• Your specialist expertise ends up embedded in the platform itself - in code, designs, and documentation - not just in advice given in meetings, and also reflected in growing knowledge and capabilities of more junior team members.
• Emerging techniques and tools get properly evaluated on their technical merits, so adoption decisions are grounded in real hands-on assessment rather than hype.
• Agents built on components you own reliably apply the correct context, explain their reasoning, justify / evidence actions, and escalate rather than guess when evidence is ambiguous or absent.

At Esko, a Veralto Company, innovation comes in every colour and never in the same package. Join Esko and see how diversity of people and thought fuels a career journey like no other. Create unique technology solutions for the packaging value chain, bring new ideas to life, make and influence decisions, and experience career growth, rewards, and recognition in our global Packaging & Colour organisations. Esko is proud to be a Product Quality & Innovation company in Veralto (NYSE: VLTO).

Imagine a world where everyone has access to clean water, safe food and medicine, and trusted essential goods. That is the tomorrow Veralto is creating today. Veralto is a $5B global leader in essential technology solutions made up of over 16,000 associates across our Water Quality and Product Identification segments all united by a powerful purpose: Safeguarding the World’s Most Vital Resources.

At Veralto, we value diversity and the existence of similarities and differences, both visible and not, found in our workforce, workplace and throughout the markets we serve. Our associates, customers and shareholders contribute unique and different perspectives as a result of these diverse attributes.