AI-Native Software Engineer (1790F8A)

Referment is working with a regulated electronic money institution that builds card issuing and embedded payments technology - the platform businesses across the UK, Europe and the US use to issue cards, manage spend and run payment programmes at scale.

The company is rebuilding its engineering organisation around an explicitly AI-assisted way of working, with executive sponsorship and real investment in the tooling and time to do it properly. This role is part of that new wave of engineers: you'll own features end to end across the full software development lifecycle, going from a rough product brief to something live in production in days, and you'll help define how the wider engineering organisation builds from here.

You'll work across a polyglot stack - Ruby, Node and Go, plus modern frontend frameworks - inside a large, complex monolith on AWS, with containers on ECS and EKS and GitHub for source control and CI. AI tooling is not an add-on here. It's core to how you build, review and QA software, how you refactor and document the codebase so AI assistants can be guided effectively, and how you turn product specifications into engineering specifications precise enough for a code assistant to deliver reliably. Delivery happens inside a regulated SDLC shaped by DORA, PCI DSS, SOC 2, EMI regulations and GDPR - change control, approvals, testing evidence and audit trails - and the expectation is that you work smart within those controls, not around them.

The Role

• Own features end to end across the full SDLC: design, build, test, deploy and support.
• Contribute confidently across Ruby, Node and Go, with frontend work in modern frameworks and web technologies.
• Use AI tooling as a core part of how you build, review and QA software, and help raise the team's AI-assisted delivery practices.
• Stay productive in a large, complex monolith - including the parts where AI tooling struggles - and use AI-assisted refactoring to improve it over time rather than work around it.
• Improve how the codebase is documented and structured so AI code assistants can be guided effectively: context files, architecture notes and well-structured inline documentation.
• Convert product specifications into precise engineering specifications with enough specificity that a code assistant delivers efficiently and reliably, with minimal iteration.
• Use proofs of concept to prove out the right level of specification and methodology so AI-assisted changes land reliably and with confidence.
• Deliver within a regulated SDLC - change management, approvals, testing evidence and audit trails - and propose ways to improve delivery speed without weakening controls.
• Build security into the SDLC by default: secure coding, threat modelling, dependency scanning, secrets and supply-chain security.
• Build and run services on AWS using containers on ECS and EKS, with GitHub for source control and CI.
• Work with data at a high standard: data analysis, ETL and data QA, supported by AI tooling.
  
  

What We're Looking For

• A proven track record as a strong engineer with genuine full-SDLC experience.
• Demonstrable evidence that your own workflow has become AI-driven - be ready to show how, not just say it.
• Solid hands-on experience in more than one of Ruby, Node and Go, and comfort picking up new languages.
• Good frontend skills and working knowledge of modern frontend frameworks and web technologies.
• Experience in large monolithic codebases, including ones AI tools find hard to navigate, with the judgement to know where AI helps and where it does not.
• A bias toward improving the codebase you inherit - including a concrete example of modernising legacy code in place.
• Hands-on experience with documentation techniques that give AI code assistants the context and guidance they need to be productive.
• Strong product sense: deciding the level of technical specificity a code assistant needs to deliver efficiently and reliably, demonstrable through practice and portfolio rather than only years of experience.
• A get-things-done operator who takes ownership and unblocks themselves, and who can absorb a complex, heavily regulated domain quickly.
• Genuine comfort operating in a regulated environment - familiarity with, or the ability to rapidly get to grips with, DORA, PCI DSS, SOC 2, EMI regulations and GDPR.
• Security built into how you develop, not just policy awareness: secure coding, threat modelling, dependency and secrets management and pipeline security, and knowing how to use AI tooling without introducing security or data-handling risk.
• Comfort across the AWS ecosystem, containerisation (ECS and EKS) and relational databases (MySQL, with some PostgreSQL).
  
  

Useful to Have

• Strength in data analysis, ETL and data QA, ideally with AI tooling.
  
  

Practicalities and Process

The role is permanent and full time, hybrid with the company's London office as the hub; remote could be considered for the right person. Visa sponsorship is not available. The selection process is deliberately rigorous and short: a written answer at application describing how AI tools are embedded in your development workflow today; a deep-dive interview into real examples from your experience; a live, screen-shared session where you extend a feature using your own AI setup and review AI-generated code; and a final conversation with the CTO, including an exercise translating a product specification into an engineering specification suitable for AI-assisted delivery.

This Could Suit...

a full-stack engineer who has genuinely rebuilt their own development workflow around AI tools and wants that to be the core of the job rather than a side experiment - for example, someone from a fintech or another regulated sector who has modernised a large monolith in place and wants to help shape how a whole engineering organisation builds with AI.

#Referment