The Apex Group was established in Bermuda in 2003 and is now one of the world’s largest fund administration and middle office solutions providers.Our business is unique in its ability to reach globally, service locally and provide cross-jurisdictional services. With our clients at the heart of everything we do, our hard-working team has successfully delivered on an unprecedented growth and transformation journey, and we are now represented by over circa 13,000 employees across 112 offices worldwide.Your career with us should reflect your energy and passion.That’s why, at Apex Group, we will do more than simply ‘empower’ you. We will work to supercharge your unique skills and experience.Take the lead and we’ll give you the support you need to be at the top of your game. And we offer you the freedom to be a positive disrupter and turn big ideas into bold, industry-changing realities.For our business, for clients, and for youAbout TocanTocan is an investor onboarding platform for the private funds market.Onboarding an investor into a private fund remains one of the most manual processes infinancial services. A single subscription can involve hundreds of pages of documentation,multiple entities across multiple jurisdictions, layered KYC and KYB requirements, and weeks ofiteration between investors, wealth managers, fund managers and fund administrators. It isslow, expensive, error prone and almost entirely people intensive.Tocan digitises that process end to end through three connected portals: a Client Portal forinvestors, a Manager Portal for fund managers and wealth managers, and an Agent Portal forfund administrators. Every party works from a single, structured, auditable view of anonboarding, from initial interest through to accepted subscription.Tocan is majority owned by a global fund administration and financial services group, which isalso our anchor client. That gives us something rare for a business of our size: direct access toinstitutional scale, real regulated volume, and a distribution route into a global client base,combined with the pace and autonomy of a small, focused product organisation.The problem is genuinely well suited to AI. Onboarding is document heavy, rules heavy,jurisdictionally complex and full of judgement currently applied manually by expert people atconsiderable cost. It is exactly the kind of domain where agentic systems can deliver materialchange, provided they are built with the rigour a regulated financial services environmentdemands.The RoleTocan is looking for a Principal Agentic AI Engineer to take technical ownership of how we buildsoftware.The immediate mandate is our agentic software delivery capability: the agents, tooling andorchestration Tocan uses across the Product Development Lifecycle. The objective is toaccelerate our progression from human led, to AI augmented, to agent led, to increasinglyautonomous product delivery. 2026 FundAdminChain Limited - Proprietary and Confidential 2 Tocan - Principal Agentic AI EngineerAlongside that, the role carries a growing responsibility for determining where and how AIshould be applied within the Tocan product itself. Delivery is the day one focus. Product is a realand near term direction rather than a distant aspiration.This is not a greenfield AI initiative. Tocan has already started the transition towards an AI nativedelivery model, and there is existing capability, tooling and agentic workflow in use today. Thesuccessful candidate will inherit that capability, form an independent technical view of it,determine what should be retained, strengthened or replaced, and take responsibility forindustrialising it into a production grade, scalable engineering capability.The role reports to the Chief Technology Officer, with scope for an expanded remit once thefoundational capability is established. The successful candidate will hire and lead two SeniorAgentic AI Engineers and will shape how that team is built.This is a deeply hands on Principal level engineering role. We are not looking for an AI strategistoperating several layers away from implementation. The successful candidate must be capableof engaging senior stakeholders, defining architecture and technical direction, and remainingdirectly involved in engineering, evaluation, testing, governance and production deployment.What Already ExistsWe would rather be specific than overstate our position, because the first thing the successfulcandidate will do is test it.Tocan has an agentic architecture in place built on Microsoft Copilot, with orchestration across anumber of other AI capabilities. Its first substantial application has been testing within theProduct Development Lifecycle, and that work is now extending into agents for UI design andfront end software engineering. Separately, Google Gemini is already used for AI augmentationwithin the Tocan platform. Software development is currently delivered through a globallydispersed capability, and this role is the beginning of building deeper engineering capability inhouse.What exists is real, working and useful. It is not yet industrialised. The evaluation, observability,governance and engineering discipline required to run agentic delivery at production scale islargely still to be built, and that is the substance of this role.Some existing approaches may need to be strengthened. Some may need to be replaced.Others will provide the foundations on which the next generation of capability is built. Thesuccessful candidate will have the autonomy and the responsibility to make those judgements.The RemitThe remit extends across the complete Product Development Lifecycle, including: Product discovery and product design 2026 FundAdminChain Limited - Proprietary and Confidential 3 Tocan - Principal Agentic AI Engineer Requirements analysis Epic, feature and user story creation Acceptance criteria Solution and technical architecture Software development Code review and refactoring Software testing and quality assurance Security and compliance Documentation CI/CD and deployment Production monitoring and support Incident analysis and remediation Continuous product improvementThe Principal Agentic AI Engineer will be expected to make the capability their own: challengingexisting approaches where necessary, establishing strong engineering and operationaldisciplines, and continually identifying frontier opportunities to change the way Tocan conceives,builds, tests and operates software. 2026 FundAdminChain Limited - Proprietary and Confidential 4 Tocan - Principal Agentic AI EngineerKey ResponsibilitiesOwn and Industrialise Tocan's Agentic CapabilityTake technical ownership of Tocan's existing Agentic AI product and software deliverycapability.Rapidly assess the existing architecture, agents, workflows, models, integrations andengineering practices.Determine what should be retained, improved, simplified or replaced.Define and implement the target architecture and evolution path for Tocan's agenticdelivery model.Move successful capabilities from individual workflows and experimental approachesinto repeatable, scalable and production grade systems.Establish the engineering standards required for reliability, security, maintainability,observability and operational resilience.Transform the Product Development LifecycleIdentify activities across Product, Design, Business Analysis, Engineering, QA andDevOps that can transition from human led execution to AI augmentation, agent leddelivery or autonomous execution.Design an agentic operating model capable of supporting work across the completeProduct Development Lifecycle.Build specialist agents capable of undertaking activities including:Product analysis and discoveryProduct and UX design supportRequirements creation and refinementEpic and user story generationAcceptance criteria creationSolution design and architectureSoftware developmentCode review and remediationTest design and test generationAutomated functional and regression testingDocumentationDeployment and release activityProduction analysis and incident remediationEnable agents to collaborate, delegate work, validate outputs and pass structuredartefacts between different stages of the delivery lifecycle.Define where human judgement, approval and oversight continue to provide genuinevalue. 2026 FundAdminChain Limited - Proprietary and Confidential 5 Tocan - Principal Agentic AI EngineerMulti-Agent Architecture and EngineeringDesign and build production grade multi agent architectures and orchestrationframeworks.Implement agent planning, reasoning, state management, memory, context engineering,tool use and inter agent communication.Build agentic workflows capable of executing complex tasks across multiple systems andstages.Develop APIs, services, connectors and tools that allow agents to interact securely withTocan's wider technology estate.Integrate agents with source control, CI/CD pipelines, issue management,documentation, testing platforms, observability tooling and cloud services.Build enterprise knowledge and RAG capabilities that allow agents to reason overproduct requirements, source code, architectural standards, business rules andorganisational knowledge.AI-Driven Software EngineeringBuild capabilities that enable agents to generate, modify, review, refactor and testproduction software.Develop autonomous and semi autonomous software engineering workflows capable ofprogressing work from requirement through implementation and validation.Introduce AI driven approaches to architecture, code review, technical debt identification,security analysis and optimisation.Integrate agentic engineering into Tocan's existing software development and releaseprocesses.Establish reusable patterns and engineering components so that new agenticcapabilities can be introduced rapidly rather than built independently each time.Evaluation, Testing and QualityEstablish systematic evaluation frameworks for agents and LLM based systems.Measure quality, reliability, accuracy, latency, cost and consistency across agenticworkflows.Implement automated evaluation, benchmarking and regression testing.Build mechanisms to detect degraded agent performance, unexpected behaviour andfailure conditions.Establish appropriate “human in the loop” controls for higher risk decisions and actions.Ensure that improvements in automation do not compromise software quality orengineering discipline. 2026 FundAdminChain Limited - Proprietary and Confidential 6 Tocan - Principal Agentic AI EngineerLLMOps, AgentOps and Production EngineeringBuild and maintain the infrastructure required to deploy, monitor and operate agenticsystems in production.Implement tracing, logging, observability, performance monitoring and evaluationpipelines.Establish versioning and deployment practices for models, prompts, tools, agents andworkflows.Optimise model selection, token consumption, latency and infrastructure cost.Design fault tolerant workflows with appropriate recovery, retry and escalationmechanisms.Governance, Security and RiskBuild appropriate guardrails, permissions, approval mechanisms and escalation pathsaround agentic systems.Ensure agentic solutions meet Tocan's security, privacy, resilience and regulatoryrequirements.Apply appropriate controls relating to GDPR, AI governance and relevant AI regulation.Ensure agent actions are traceable, auditable and appropriately permissioned.Work with internal stakeholders to ensure that increasing levels of autonomy areintroduced safely and deliberately.Build and Lead the TeamRecruit two Senior Agentic AI Engineers and establish how the team works.Set the engineering standards, review practices and technical direction the teamoperates to.Develop capability so that agentic engineering becomes a shared discipline rather thanan individual one.Set technical standards and patterns that apply across Tocan's globally dispersedsoftware development capability.Remain personally hands on. This is a building role, not a management role.Shape the Application of AI Within the ProductHelp determine where AI and agentic capability can deliver genuine value within theTocan platform. 2026 FundAdminChain Limited - Proprietary and Confidential 7 Tocan - Principal Agentic AI EngineerAssess candidate use cases across investor onboarding, document handling, dataextraction and validation, and operational workflow.Advise on architecture, feasibility, cost, risk and regulatory considerations for productfacing AI.Ensure product facing AI is held to the same standards of evaluation, observability andgovernance as internal agentic systems.Delivery is the day one priority. This strand grows as the foundational capability matures.Drive Frontier ChangeMaintain a strong understanding of frontier developments across Generative AI, AgenticAI, coding agents, reasoning models, orchestration and AI native software engineering.Identify emerging capabilities that could materially improve Tocan's product andengineering model.Rapidly prototype promising approaches and move successful capabilities intoproduction.Challenge established product and engineering practices where better agenticapproaches are available.Continuously raise the level of autonomy achievable across the Product DevelopmentLifecycle.Demonstrate measurable improvements in engineering throughput, delivery cycle time,quality and cost.Skills Required and Level of ExperienceExperienceThe successful candidate will bring significant senior software engineering and architectureexperience combined with demonstrable recent experience delivering production AI systems.We would expect:10+ years across software engineering, architecture, platform engineering or enterprisetechnology delivery. Exceptional candidates are likely to bring 15+.Significant experience operating at Principal, Staff, Lead Architect or equivalent technicallevel.Recent, demonstrable experience designing and delivering production Generative AIand Agentic AI systems. We mean systems real users depend on, not proofs of concept.Proven hands on experience designing and building multi agent architectures and agentorchestration. 2026 FundAdminChain Limited - Proprietary and Confidential 8 Tocan - Principal Agentic AI EngineerExperience taking AI capability from architecture and design through engineering,evaluation, governance, deployment and production operation.Strong production software engineering capability, with depth in at least one moderntyped language such as Kotlin, Java or TypeScript, alongside the Python or TypeScriptproficiency required to build agentic tooling.Experience applying AI to software engineering, product delivery or wider SDLCautomation.Experience operating within complex enterprise organisations, major technologyconsultancies, high growth technology businesses or technically demanding start ups.Evidence of personally building and operating systems, rather than managing AIprogrammes or advising delivery teams.What We Will Ask You AboutFamiliarity with the vocabulary of this field is now widespread. We are more interested inspecifics, and the selection process is built around them. Expect to discuss:An agentic system you personally designed and took to production: what it did, what itcost to run, and how you knew it was working.How you evaluated it, and what your evaluation actually measured.How it failed, what that taught you, and what you changed as a result.A point at which you concluded an AI approach was the wrong answer, and what you didinstead.How you decided where a human needed to remain in the loop, and how you tested thatjudgement.Technical CapabilityStrong practical experience is expected across a number of the following:Agentic AI, multi agent architectures and agent orchestrationAgent planning, reasoning, memory, state management and context engineeringLLM tool and function calling, and structured outputsEnterprise RAG and knowledge systemsLLM and agent evaluationAI assisted and AI driven software engineeringPython or TypeScript for agentic tooling, and strong production engineering in at leastone of Kotlin/JVM or TypeScriptCI/CD and automated software delivery, including test automation across web front endsand JVM servicesObservability, distributed tracing, LLMOps and AgentOpsAI security and guardrails 2026 FundAdminChain Limited - Proprietary and Confidential 9 Tocan - Principal Agentic AI EngineerCloud native architecture on Azure, including Azure OpenAI or AI Foundry, Cosmos DBand the wider Azure estateOur current estate includes Microsoft Copilot and Google Gemini, and we evaluate frontiermodels, coding agents and orchestration frameworks continuously. We are considerably moreinterested in depth of engineering capability, architectural judgement and demonstrableproduction delivery than in experience with any particular model or framework.Tocan's Technology EstateThe Tocan platform is built in Kotlin with a React and Ant Design front end, running on Azurewith MySQL and Cosmos DB. Agentic capability must therefore be designed to operate againstthis estate, not in isolation from it. We do not expect deep prior experience with every element,but the successful candidate must be capable of forming a credible engineering view of thecode that agents will generate, review, test and deploy.The technology landscape is evolving rapidly. The successful candidate must be capable ofevaluating emerging technologies objectively and adopting, replacing or discarding toolingbased on measurable engineering outcomes rather than vendor or framework preference.Personal AttributesFundamentally an engineer and builder, not solely a strategist.Comfortable moving between architecture level thinking and hands on implementation.Confident enough to inherit an existing capability and form an independent technicalview of it.Willing to challenge or replace existing approaches where a better solution exists.Pragmatic about both the opportunities and the limitations of current AI technology.Understands that production Agentic AI requires robust engineering, evaluation,observability and governance.Able to explain complex technical concepts to senior business and technologystakeholders.Operates effectively with significant autonomy and ownership.Motivated by measurable business and engineering outcomes rather thanexperimentation for its own sake.Curious, technically ambitious and motivated by pushing the frontier of how softwareproducts are delivered.What Success Looks LikeThe First Twelve Months 2026 FundAdminChain Limited - Proprietary and Confidential 10 Tocan - Principal Agentic AI EngineerSuccess in the first twelve months means a material transition to an agentic productdevelopment pattern, evidenced across four areas:UI designFront end engineeringTestingDeploymentIn each, we expect a demonstrable shift from human led execution to agent led delivery withhuman oversight, supported by the evaluation, observability and governance required to trust it.Alongside that: two Senior Agentic AI Engineers recruited and productive, and a clear, credibletechnical roadmap for extending the pattern further across the Product Development Lifecycle.Beyond the First YearSuccess will ultimately be measured by how materially Agentic AI changes Tocan's ability to: Conceive and design product Translate product intent into actionable requirements Build software Test and assure software quality Release product more rapidly Operate and improve production systems Increase engineering throughput and reduce delivery cycle time Improve quality and automated test coverage Improve the economics of software delivery Increase the proportion of the Product Development Lifecycle undertakenautonomously Achieve all of this without compromising security, governance, resilience or regulatorycompliancePractical Details Location: London Working pattern: Hybrid Employment type: Permanent Reporting line: Chief Technology Officer Team: two Senior Agentic AI Engineers, recruited by the successful candidate Remuneration: competitive, and discussed openly at an early stage 2026 FundAdminChain Limited - Proprietary and Confidential 11 Tocan - Principal Agentic AI EngineerApplicants must have the right to work in the United Kingdom. Tocan is an equal opportunitiesemployer and welcomes applications from all backgrounds. Personal data submitted as part ofan application is processed solely for recruitment purposes and handled in accordance withapplicable data protection law.A Final NoteTocan is already on this journey. We are not recruiting someone to write an Agentic AI strategyor to begin with a blank sheet of paper. Existing capability, technology and expertise are alreadyin place.We are looking for an engineer who has personally built and operated production AI systems,who will step into that environment, develop their own technical view, and take ownership ofwhat comes next.Our ambition is an AI native product and engineering organisation in which increasinglysophisticated activities across Product, Design, Engineering, Testing and Operations areundertaken by autonomous or semi autonomous software agents, with people focused wherehuman judgement, creativity and accountability provide the greatest value.The Principal Agentic AI Engineer will play a central role in determining how far and how quicklyTocan can push that frontier.Disclaimer : Unsolicited CVs sent to Apex (Talent Acquisition Team or Hiring Managers) by recruitment agencies will not be accepted for this position. Apex operates a direct sourcing model and where agency assistance is required, the Talent Acquisition team will engage directly with our exclusive recruitment partners.
Principal Software Engineer in London employer: Apex Group Ltd
Apex Group is an exceptional employer, offering a dynamic work culture that fosters innovation and collaboration in the financial sector. With a strong focus on employee growth, we provide ample opportunities for professional development and career advancement, all while working in a vibrant multinational environment that values diversity and inclusion. Join us to make a meaningful impact in global treasury and funding strategies, backed by a supportive team and a commitment to excellence.