Principal Software Engineer
Most senior hands-on engineer at Defaqto (Fintel), owning engineering standards and leading the shift from AI-assisted to AI-led delivery, plus direct ownership of the highest-risk initiatives such as developer- and AI-agent-facing APIs and MCP servers. Core stack: .NET/C#, TypeScript on Azure, Python, PHP and SQL.
- Compensation
- Not disclosed
- Location
- Farringdon, London, London
- Work setup
- Hybrid
- Seniority
- Principal
- Employment
- Contract
- Experience
- Not specified
- Category
- Software engineering
- Posted
Skills
Must have
- Substantial commercial software engineering experience, including staff/principal level, with technical influence beyond own team
- Deep, current, hands-on ability in a modern stack (.NET/C# and TypeScript on Azure, Python, PHP, SQL and modern data tooling)
- Evidenced experience leading agentic and AI-assisted development in a production engineering organisation, with measurable outcomes
- Clear point of view on AI-led delivery practices: specification and context engineering, test-first contracts, agent-readable codebases, review as first-class discipline
- Demonstrable ownership of engineering standards and practices across multiple teams, embedded in tooling
- Strong API and integration engineering: designing, versioning, securing and operating interfaces consumed by other systems and their agents
- Treats data governance as an engineering problem: entitlements, licensing, lineage, provenance, auditability
- Sound distributed systems and cloud architecture judgement (resilience, observability, performance, cost, operability)
- Excellent influencing and communication skills across product, commercial and executive audiences
- Commercial instinct tying technical work to measurable business outcomes
- Mentoring track record making senior engineers materially better
- Comfort with a moving remit and owning whichever problem is hardest right now
- Legal right to work in the UK without time restrictions or future sponsorship
Nice to have
- Experience building or operating agent-facing integration surfaces such as MCP servers or tool-calling interfaces
- Financial services, fintech or regulated data experience; comfort treating FCA and UK GDPR/ICO expectations as design constraints
- Experience with data products, business intelligence or analytics platforms (data modelling, contracts, quality, lineage, entitlements)
- Experience of a developer portal, API marketplace or platform-as-a-product
- Platform engineering and developer experience: CI/CD, internal developer platforms, golden paths, policy-as-code, self-service tooling
- Event-driven architecture and experience of legacy modernisation or migration programmes
- Contribution to the wider engineering community: open source, writing, speaking or standards work
Full 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.
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.