The job
Principal Software Engineer
Department: Technology
Employment Type: Permanent - Full Time
Location: London
Reporting To: Head of Engineering
Description
As a Principal Software Engineer, you'll be the most senior individual contributor in our Technology function. Reporting to the Head of Engineering and working closely with the CTO, you'll own our engineering standards, technical direction, and above all - our shift from AI-assisted coding to AI-led delivery, where engineers direct and verify rather than type.
This is a hands-on technical leadership role with no requirement to manage people, unless you want to. You'll guide several squads, earning influence through judgement and delivered outcomes, while taking direct ownership of whichever initiatives carry the most technical risk or upside at any given time. Our current focus is opening Defaqto data to developers and AI agents through well-governed APIs and MCP servers - but the role is defined by the level you operate at, not by any one product, so expect what you own to change as our priorities do.
You'll also be a genuine sounding board for technology leadership, equally comfortable shaping multi-year strategy or unblocking a production issue this afternoon. If you're the engineer others go to when a decision really matters, and you want the remit to raise the bar across the whole organisation, this is that role.
What you'll do
Lead our shift from AI-assisted coding to AI-led delivery
- Own the trajectory from engineers using assistants to write code, to engineers directing agents that write it - with humans firmly in the loop as architect, reviewer and accountable owner.
- Define what good looks like at each step: prompt and context engineering, agent-readable codebases (clear structure, strong typing, rich tests, machine-readable specs), and tooling that makes agent output reproducible, not lucky.
- Make review the new craft - tests as the contract, static analysis and security scanning built into the loop, and no unreviewed AI-authored change reaching production.
- Shift practices towards specification over syntax, so engineers focus on problem framing, acceptance criteria and verification rather than implementation keystrokes.
- Run structured experiments with new models and workflows, measuring lead time, change failure rate and developer experience - retiring anything that doesn't earn its place.
- Set the guardrails that make this safe in a regulated business: secure, licence-aware tooling, provenance and auditability for generated code, and a named human accountable for everything that ships.
- Coach engineers through the shift from writing code to engineering outcomes, bringing sceptics along with enthusiasts.
Turn developer practices into engineering practices
- Build engineering discipline that holds at organisational scale and at the volume agentic delivery produces.
- Codify practice into the paved road - templates, pipelines, linters, policy-as-code and agent instructions - so the right thing is the easy thing.
- Define and evolve engineering standards across Defaqto Technology: code review, testing, observability, security by design and definition of done.
- Hold the line on quality through coaching and pairing rather than policing, and help raise the hiring bar to match.
Take hands-on ownership of what matters most
- Own the highest-risk, highest-upside initiatives end to end - architecture, build, launch and operational reality. What that is will change; the level you operate at won't.
- Design platforms and integration surfaces for both human developers and AI agents: machine-readable capability discovery, unambiguous tool definitions, MCP servers, predictable errors and sensible rate limiting under agentic access patterns.
- Build data governance into the technology itself - entitlements, licensing, lineage and provenance enforced at the point of access, with full audit of who or what accessed data, and when.
- Make quality and trust observable through data contracts and clear freshness signals, so no consumer can silently misread a rating or field.
- Own the consumption experience end to end - onboarding, SDKs, documentation, versioning - treating time to first successful call as a metric you own.
Partner with Product
- Pressure-test feasibility, cost and risk on product ideas early, before commitments are made.
- Surface technology-led opportunities Product hadn't thought to ask for.
- Translate between commercial intent and technical reality in both directions.
Guide squads and evolve the tech stack
- Guide teams towards sound, well-supported technology choices over novelty for its own sake.
- Own continuous improvement of the stack, with every material change carrying a stated benefit and migration plan.
- Lead cross-team architecture decisions, document them as ADRs, and keep our technical debt position honest.
Be a strategic and tactical sounding board
- Act as trusted counsel to the Head of Engineering, CTO and leadership on both strategy and this week's trade-offs.
- Represent engineering reality in commercial, regulatory and roadmap conversations.
- Stay hands-on, building and reviewing the work that matters most.
What you'll need to succeed:
Essential requirements:
- Substantial commercial software engineering experience, including time at staff/principal level, with technical influence beyond your own team.
- Deep, current, hands-on ability in a modern stack, with breadth beyond your strongest language. Our core stack: .NET/C# and TypeScript on Azure, Python, PHP, SQL and modern data tooling.
- Real, evidenced experience leading agentic and AI-assisted development in a production engineering organisation - not just personal assistant use. You can show how you changed a team's practices and kept quality intact while doing so.
- A clear point of view on what makes AI-led delivery work: specification engineering, test-first contracts, agent-readable codebases, and review as a first-class discipline.
- Demonstrable ownership of engineering standards across multiple teams - written, embedded in tooling, and proven to change behaviour.
- Strong API and integration engineering, with real care for the experience of consumers - human and increasingly agentic.
- Treats data governance as an engineering problem, enforced in the technology rather than a policy document.
- Sound distributed systems and cloud architecture judgement - resilience, observability, performance and cost, not just diagrams.
- Excellent influencing and communication skills across product, commercial and executive audiences.
- Commercial instinct - you tie technical work to outcomes the business can measure.
- A mentoring track record: senior engineers have visibly grown through working with you.
- Comfort with a moving remit - motivated by the hardest problem right now, not by owning one system indefinitely.
Desirable requirements:
- Experience building or operating agent-facing integration surfaces - MCP servers, tool-calling interfaces or similar.
- Financial services, fintech or regulated data experience, comfortable treating FCA and UK GDPR/ICO expectations as design constraints.
- Experience with data products, BI or analytics platforms — modelling, contracts, quality and lineage.
- Experience of a developer portal, API marketplace or platform-as-a-product.
- Platform engineering and DX work: CI/CD, internal developer platforms, golden paths, self-service tooling.