Senior AI Engineer - AI Transformation

About the role

You own the design and delivery of end-to-end agent systems that redefine how Multiverse operates and serves learners across the UK and Europe. You define context and retrieval strategies that balance cost, quality, and latency to ensure AI behaviour remains accurate, safe, and helpful for every stakeholder. You build and own evaluation frameworks that turn vague ideas about AI quality into measurable, automated signals and human-reviewed insights. You design robust tool integrations and data contracts so that autonomous agents can reliably interact with existing products, internal tools, and customer-facing systems. You influence technical direction through evidence-based arguments, sharp architectural thinking, and a willingness to challenge the team when the path is unclear. You raise the bar for everyone by conducting rigorous code reviews, pairing on hard problems, and modelling what production-grade AI engineering looks like in practice. You use Claude Code as your core development environment, shaping context, constraints, and review loops to get the most reliable output from AI-assisted workflows. Your work inside Multiverse becomes the blueprint that other organisations will copy when they rebuild their own teams for the AI era.

Key facts

What you'll do

Design and deliver complete agent systems that solve real product problems, owning architecture, implementation, evaluation, and production operation from idea to live service.

Define context and retrieval strategies, deciding what sits in the context window and what lives in external stores, while managing trade-offs between cost, quality, speed, and safety.

Build evaluation frameworks that combine automated metrics for accuracy, safety, helpfulness, domain-specific quality, and latency with human-in-the-loop review processes that feed directly into product decisions.

Design and implement tool integrations including MCPs, APIs, data contracts, and error handling so that autonomous agents can interact reliably with internal and external systems.

Influence technical direction by forming and backing opinions about architecture, pushing back when risks appear, and proposing better alternatives that align with product and business goals.

Review code with rigour and provide feedback that improves quality and capability across the team, while pairing with less experienced engineers on difficult problems to accelerate their growth.

Use Claude Code as the primary development workflow, setting context, defining constraints, reviewing outputs critically, and augmenting the tool with domain knowledge and engineering discipline.

Establish operational practices that keep AI systems performant and reliable in production, including monitoring, debugging, and continuous improvement of agent behaviour over time.

Collaborate closely with the wider engineering organisation building Multiverse's customer-facing product to ensure agent workflows integrate cleanly with existing services and data models.

Translate ambiguous product goals into concrete technical milestones, prioritising work that de-risks the most important assumptions about AI behaviour, reliability, and user impact.

Requirements

You have a strong background in software engineering and a deep understanding of how AI systems work in practice, not just in theory.

You are experienced in building and operating production-grade AI applications, with a track record of delivering agentic systems that work reliably at scale.

You understand context engineering and retrieval design deeply, including token efficiency, relevance scoring, and the impact of chunking strategies on downstream model performance.

You have hands-on experience with evaluation techniques for AI systems, including automated metrics, human review processes, and safety guardrails.

You are fluent in tool integration patterns such as MCPs, API design, data contracts, and error handling strategies for distributed systems.

You have worked with large language models and are comfortable making architectural decisions that balance performance, cost, and maintainability.

You are comfortable working across team boundaries, integrating with existing products and legacy systems while helping modernise them in a responsible way.

You communicate clearly and collaborate effectively, able to translate technical trade-offs into decisions that non-technical stakeholders can understand and act on.

Nice to have

Experience contributing to open source AI projects or publishing technical work that demonstrates your understanding of agent systems.

Familiarity with the specific tools and frameworks Multiverse uses for data, learning delivery, and workforce analytics.

Practical notes

This is a full-time role based in London.

The position requires close collaboration across engineering, product, and operations teams with limited bureaucratic overhead.

You must be comfortable working within existing operational constraints while driving change that scales across the business.