Principal Engineer

CreateFuture is fast becoming the UK’s most recognisable digital consultancy, with years of experience building digital products and services for major organisations whilst putting our people first. We have offices in the centre of Edinburgh, Leeds, Manchester and London as well as remote employees located throughout the country. 

We are a team of creators - whether that’s code, project plans, go to market strategies, culture initiatives, marketing campaigns, large language models or people policies. And together, with our clients, we create the future. This has seen us collaborate and partner across a multitude of industries and sectors, with the likes of PayPal, Adidas, NatWest, FanDuel and Money Saving Expert to name just a few.

Our reputation as a partner determined to deliver high-quality, robust and thoughtful products has enabled us to scale to over 500 people in the last couple of years, and it is our amazing people - along with the safe, supportive and friendly culture we have built - that makes CreateFuture a great place to work. Don’t just take our word for it though, we have been recognised by Best Workplaces UK multiple years in a row - across a number of categories - and our employee exit rate is astonishingly low.
Join us on our journey… Let’s create something awesome, together, today.

The bar has moved. Five years ago a strong engineer was measured by what they could build and how well they could mentor others to build it. Today, with Claude Code and agents, we produce quality work at a speed that would have looked superhuman in 2022.

What separates exceptional from good is no longer typing speed or local craftsmanship. It is the complete package: architecture, deep engineering foundations, communication, system thinking, and the judgement to direct both your own intelligence and AI toward outcomes that matter. 

Exceptional people at this level are not the best implementer in the room; they are the person whose specifications, decisions, and review judgement compound into systems that work in production and that others (and agents) can build on. They are a force multiplier, they hold a room at the highest levels, and clients treat them as the SME by default.

The mindset is the engine, and it is almost impossible to teach retrospectively. You are outcome obsessed, not output obsessed and you care whether the problem was actually solved - you resist the ego instinct to admire a clever solution and ask whether it was needed at all.

You are intellectually honest about uncertainty, taking a trust-but-verify approach and assuming the surprising bug is the one that will burn. You are deeply pragmatic, which matters at CreateFuture where pace counts: you have outgrown the religious wars about platforms, languages and patterns, you hold opinions loosely, and you decide for the systems in front of you, not the ones you wish you had.

You have a bias toward delivery: the smallest thing that works, then iterate, treating documents and meetings as means to that end. 

And you are relentlessly curious beyond any immediate need, wanting to know how the database actually stores things or why a Claude Opus inference call costs what it does, which keeps you several layers ahead while the rest of the team is still reading the API docs.

For a Principal to be truly exceptional, this is now already a first-class competency in your arsenal, not a side skill.  An engineer who is brilliant by last year's standards but not fluent with AI is not exceptional in 2026. In practice that means:

• Setting AI up to succeed. You treat context engineering as a craft (the right files, specs, examples and constraints), and you architect systems with clean interfaces, good naming, machine-readable specs and knowledge bases so the codebase is a quality baseline AI can work on later. No separate ivory-tower architect required.
  
• Reviewing and choosing well. You read far more code than you write, and your value is in the quality of review: catching the silent assumption, the security flaw, the architectural drift, knowing where models fail and looking there first. You pick the right tool for the job (Claude Code for a deep refactor, an inline assistant when that is enough, a sub-agent for parallel work, the right model for analysis versus generation) and you know the constraints differ by client, an enterprise like PayPoint or Natwest versus a scale-up type business like Zuto or OAG.
  
• Working with non-determinism. You see where AI moves the bottleneck and help people left and right of you unblock the value stream. You build with AI components but know success is no longer a binary, reaching for evals, golden datasets, structured outputs and fallback paths as naturally as you reach for unit tests.

Exceptional means depth in at least one area but competence everywhere the system reaches.

• Backend: an instinct for distributed systems (what fails, what is eventually consistent, where queues back up), APIs that age well, thoughtful data modelling, and an understanding that languages are mostly syntax differences you can close with AI.
  
• Cloud: cost-aware and security-aware by default, reaching for IaC, composing the primitives rather than stitching together managed services and hoping it works.
  
• Data: modern pipelines, governance and lineage, increasingly vector stores and embeddings, with data quality treated as seriously as production code because garbage in, garbage out is far worse once AI features depend on it.
  
• Web and mobile: past the framework arguments, with a clear view on what you would pick today, and able to build "good enough" front ends yourself with AI.

Full-stack here means seeing the platform as one system and knowing which levers matter most. It does not mean front-end work with a few Node.js servers.

At CreateFuture we work in a consultancy context, where many otherwise excellent IC engineers can struggle. 

You know technology is roughly 30% of the picture; the rest is design, argument and reasoning, commercial and political acumen, and the clarity to frame outcomes for a COO, CTO, CIO or CEO in their own language.

You know when gold-plating is wasteful, you can name when a project's real problem is not technical and say so even when it is awkward, and you have a strong sense of what is durable versus what is fashion. 

You are not the engineer who introduced the team to seven new technologies last year; you are the one who removed three and quietly made the remaining stack stronger.

Technical work is necessary but no longer sufficient. By default you raise the level of everyone around you, regardless of discipline.

• Communicate clearly and frequently (documents, decisions, post-mortems, design notes); the written artefact is the machine-readable context we can feed to AI later.
• Make decisions and own them. You don't escalate everything or hoard control, and you are willing to be the benevolent dictator when it is your turn.
• Mentor in AI fluency, not just code: tooling, prompting and review discipline, so your team is twice as productive a year from now.
• Need little line management. You communicate upwards, raise by exception, self-resolve, and can be trusted with financial controls.
• Disagree well. You tell people they are wrong with evidence and no ego, and you change your own mind when you are shown to be wrong.

• The gifted engineer who sits in a silo. This is a team game.
• The one who is fluent with AI but ships without the guardrails and quality gates enterprise clients expect.
• The one with strong technical opinions who cannot justify them commercially and gets defensive when challenged.
• The pure manager who wants to orchestrate teams rather than get involved. This is about setting a new standard, not coordinating it.
• The engineer who will not immerse in a client environment. This role will mean on-site work and trusted relationships built in person. We are not going back to working away from the client and being bypassed.