Independent software engineering / London

AI powered by experience.

I’m David Naylor, an independent software engineer. I design and build production-grade business systems that meet today’s demanding standards. More than twenty years of hard-won experience lets me harness AI to deliver above and beyond what’s ever been possible, and to help your organisation do the same.

Experience includes

  • Checkout.com
  • ClearBank
  • Just Eat
  • RELX
  • Toyota (GB)

01 / The shift

AI is the lever. Experience knows how to use it.

Generative AI is a lever: code is now cheap to produce in any quantity, good or bad. Experience brings the skill and judgement to harness it, knowing where to apply it and when to trust what comes back. Research on teams using AI keeps finding the same thing: the tools multiply the skill and discipline that are already there.

Google DORA / 2025 report

An amplifier

AI doesn’t fix a team; it amplifies what’s already there. Strong teams use AI to become even better and more efficient.

State of AI-assisted Software Development

Fastly developer survey / 2025

Experience decides

About a third of developers with ten or more years’ experience say over half the code they ship is now AI-generated, and 59% say AI tools help them ship faster. Experience, not the tools themselves, appears to be what determines who actually benefits.

Senior developers ship more AI code

Stack Overflow Developer Survey / 2026

Trust, once verified

Among developers with sixteen or more years’ experience, 69% view AI tools favourably. Trust is conditional: the most common answer, at 48%, was I trust it when I can easily verify the output.

2026 Developer Survey: AI

What agents bring

  • Speed, at every stage of delivery
  • Breadth: more options explored before committing
  • Code, tests and documentation in any volume
  • A second look at anything, on demand

What experience brings

  • Direction: an engineer in the driving seat
  • Judgement: telling good from bad
  • Architecture that holds up under load
  • Accountability for what goes live

02 / Services

Three ways to work together.

The same practice at three levels: building a system, leading the work, or bringing this way of working to your team. In each, experience and agents work together, and I work with your team, inside it or beside it. Engagements run through Bitform Ltd on a contract basis.

01

Hands-on delivery

I design and build backend platforms, APIs and the services behind them, embedded in your team or alongside it. My strength is systems that must stay correct under load: high-throughput APIs, event-driven services and distributed data. I work with agents at every stage, so the work arrives sooner and to a higher standard, and I answer for the design and the result.

  • Backend platforms and APIs
  • Event-driven systems
  • Distributed data
  • Cloud and DevOps

02

Technical leadership

Architecture and technical direction for a system, a migration or a programme of work, and leadership of the team that delivers it. I set the design, the standards and the checks within which engineers and agents can move quickly and still be trusted. The faster code can be produced, the more its direction matters.

  • Architecture
  • Team leadership
  • Live migrations
  • Performance at scale

03

Agentic engineering for teams

I bring this way of working to your engineers, so they gain the speed of agents without losing control of quality. I introduce it on live work, alongside the team: spec-driven development, shared skills, engineering standards and automated evals, built into its own repositories. The practice stays when I leave.

  • Spec-driven development
  • Shared skills and standards
  • Automated evals
  • Frontier and local models

03 / Approach

Engineer and agent, better together.

Every stage of delivery is a partnership, led by the engineer. The agent brings speed and breadth: drafts, options and checks on demand. The engineer brings direction and judgement, and decides what stands. They work in rounds, each building on the other’s last, and the outcome is elevated: work delivered sooner, and to a higher standard, than either would manage alone.

That holds only where the work is checked. DORA’s research found that AI adoption still goes with lower delivery stability unless testing, version control and feedback loops are strong. So when code is cheap, verification is the work, and it runs through every stage below.

  1. 01

    Specify

    Requirements are worked into a written spec over several rounds: the agent drafts and probes for gaps, the engineer decides. The spec gives every later agent its context and scope, and it defines what correct means, with guardrails and tests in place before any code.

  2. 02

    Plan

    The spec is distilled into stories, each small enough to verify and each with its acceptance criteria written down. Agents propose the breakdown; engineers settle what is in, what is out and what comes first.

  3. 03

    Build

    Stories become tasks, and agents implement them inside the guardrails. Every task is a pull request, reviewed by a second engineer who works with agents too and decides how much of the review to automate. Nothing merges on trust.

  4. 04

    Ship

    Small releases, with monitoring and a way back. Agents assist here as well, with pipelines, infrastructure code, release notes, runbooks and incident diagnosis. People stay accountable for what goes live.

04 / Experience

Twenty years of systems that had to work.

I’ve worked as an engineer, team lead, engineering manager and founder, at FTSE 100 companies, scale-ups and start-ups.

20+
years building production software
300
payments a second at peak, on a gateway I led
10×
throughput gain on a bank’s core payment system
1M+
registered users on a platform whose backend I led from pre-launch
  1. 2023 – now

    Toyota (GB)

    Designed and built core services on the platform that takes Toyota and Lexus used vehicles from intake to sale across the UK, spanning four systems and more than twenty integrations. Led the service that brings dealers’ used cars to toyota.co.uk from first discovery to live operation, and created the agentic development framework that the engineering team now builds with across eight repositories.

  2. 2020 – 2023

    Checkout.com

    Managed the nine-strong engineering team behind the payment gateway of Checkout.com, which became the UK’s most valuable private fintech in that time. The gateway is the public API through which its merchants take live payments, at up to 300 transactions a second. Conceived and delivered the migration of the gateway’s primary database to DynamoDB while it remained in full service, raising availability and removing substantial licensing and infrastructure cost.

  3. 2019 – 2020

    ClearBank

    Led the development of the high-performance core that manages live payment transactions and account balances for ClearBank, as the bank moved from a monolith to event-driven microservices on Azure. Rethinking how balances are stored and updated delivered a tenfold increase in throughput, on significantly less infrastructure.

  4. 2017 – 2019

    Just Eat FTSE 100

    Strengthened and scaled the global Orderpad platform through a period of rapid growth, as it carried orders to around 80,000 restaurants worldwide at peaks approaching 70 a second. Resolved the bottlenecks and race conditions that surface only at that scale, and launched the platform in new countries and regions on shared, multi-tenant infrastructure that reduced its running costs.

  5. 2017

    Reed Exhibitions, RELX FTSE 100

    Advanced the recommendation engine that connects each trade-show visitor with the exhibitors most relevant to them. Designed the Elasticsearch models and queries at the heart of its recommendations and, working with the group’s data science team, extended it to draw on new sources of data.

  6. 2012 – 2014

    Bloom.fm

    Led the backend engineering of a music streaming service from pre-launch through rapid growth to more than a million registered users, with a catalogue of 22 million tracks. Delivered search across the entire catalogue, more than 150 radio stations curated by machine learning, and a subscription platform spanning iTunes, Google Play and card payments.

  7. 2002 – 2017

    Earlier

    Led the creation of real-time chat and search for the live-video platform Kastr. Founded Appakin, a search engine that used machine learning to organise apps into more than a thousand categories. Co-founded the consultancy Altitude Business Systems and grew it over five years, leading the team behind the multi-award-winning website and reservation system of the airline Silverjet.

Education

  • MSc Embedded Systems and Robotics, Distinction University of Essex
  • BSc (Hons) Computer Science University of Pretoria

Specialisms

  • Backend platforms and APIs C# and .NET in depth: design, architecture, performance at scale
  • Cloud and DevOps Cloud-native architecture, infrastructure as code and CI/CD
  • Event-driven systems Microservices, queues, brokers and event streams
  • Distributed data Relational and NoSQL databases, search and caching
  • Agentic engineering Coding agents on frontier and local models

05 / Contact

Tell me what you’re building.

A few lines about the problem, the team and the timescale is enough to start. I read and answer every message myself.