Work history
ClearSignal (opens in new tab)
Head of Engineering
I lead engineering for ClearSignal, a B2B platform for continuous AML, KYC and KYB compliance and credit decisioning. I set the product and technical direction, wrote the first version myself, and took it from an empty repo to general release in 18 months.
What I built:
- Product architecture from the ground up
- Real-time risk pipeline aggregating Companies House, HMRC, OFSI, UN, CCJ, PEP and adverse-media sources, surfacing live AML, KYC and KYB events to compliance teams as they happen
- Multi-LLM layer (OpenAI, Gemini Enterprise, Copilot SDK, FastMCP) with retrieval over trusted data, turning raw signals into explainable, auditable decision support
- AI policy-to-rules engine, credit decisioning, and RAG over regulatory guidance, policy, and case precedent
- Hybrid on-prem/cloud platform: a .NET Core backend on IIS with on-premise SQL Server, plus a Google Cloud tier running Dockerised services on Kubernetes (GKE), Cloud SQL, BigQuery for risk analytics and VPC networking, with Hangfire driving the continuous monitoring jobs
- Vue.js front end with an evidence-backed audit UI, so compliance teams act with a documented rationale
- Shipped across four surfaces: web app, public API, Salesforce integration, and an MCP server for clients' own AI agents
- CI/CD pipelines and deployment infrastructure
I wrote the v1 codebase, built and lead a Cardiff engineering team of five, and partnered directly with clients across legal, finance and real estate to shape the product, onboarding five firms since launch.
Your People
Head of Engineering
Re-architected a multi-tenant SaaS platform. The new system ran side by side with the old one on shared data, so existing customers kept working while we migrated them across without disruption.
What I did:
- Ran the engineering function for a live multi-tenant SaaS, leading a team of 25 and keeping delivery aligned with product and commercial priorities
- Rebuilt the platform from scratch alongside the legacy system, sharing the same data, then ran a safe, incremental migration for live customers
- Designed the core architecture, data model, and component flows for the new platform
- Built AI-assisted decision engines using LLMs and RAG, drawing context from membership profiles and company knowledge bases to sharpen workflow recommendations
- Reviewed the legacy platform and sequenced the weaknesses worth fixing first, rather than risking a big-bang rewrite
- Cut cloud and infrastructure costs by 40% through architectural change
- Kept shipping features on both the legacy and new platforms throughout the transition
The hard part wasn't the rebuild. It was keeping every existing customer live while it happened.
Blue Tahiti Software Ltd
Head of Engineering
Built and shipped SaaS products from zero to paying customers. Designed and built early MVPs and technical stacks, set up infrastructure, hired teams, and supported live customers in production.
This is where I learned how to build real products, not just prototypes, but systems companies paid for and depended on.
What I did:
- Took multiple SaaS products from idea to production across architecture, code, and deployment
- Built and grew engineering teams to 30, hiring for quality and ownership and staying in the code as a player-coach
- Chose the stack and infrastructure for each product, then designed and ran the networks and servers behind them
- Set up how the engineering and product teams worked day to day
- Worked with the commercial side on positioning, branding, and go-to-market
- Supported live customers across SME and enterprise accounts
- Ran production infrastructure across Google Cloud and Azure: VPC networking, IAM and security, containerised workloads on Kubernetes (GKE and AKS), and managed Cloud SQL databases
Earlier roles
Top Banana Computers Ltd (OP3.co.uk)
Technical Lead (Apr 2009 – May 2010)
Technical lead on a data-heavy web application, building database schemas, application logic, and UI. This was my first role owning the full technical stack on a commercial product.
Lloyds TSB Asset Finance Division
Systems development and support (Sep 2004 – Apr 2009)
Built and maintained systems in a regulated bank using SQL and C#, at enterprise scale. This is where I learned how compliance actually works and why you can't ship software that breaks in a bank.
Virgin Media
Technical Support Advisor (Sep 2003 – Sep 2004)
Front-line technical support for consumers and businesses, helping customers fix internet and network problems. I started here right out of school and learned how systems really work by fixing them when they break.
Things I've built
AI products I've taken from idea to production, mostly as sole engineer.
MarketStrike (opens in new tab)
AI-powered B2B sales intelligence, built end to end as sole engineer
An AI sales-intelligence platform that surfaces the accounts worth acting on now and gets evidence-backed outreach ready for review. Taken from zero to production.
- "TripWire", a real-time market-signal pipeline: ingestion via Pub/Sub, Python AI agents (Claude, Gemini, Copilot SDK), and actionable alerting
- Autonomous agents for lead research and ICP generation, built on FastMCP and the Claude and Gemini SDKs
- Full stack: Vue/Nuxt front end, .NET Core and Python back ends, PostgreSQL, Docker and Kubernetes on Google Cloud Run, Cloudflare Workers, and GitHub CI/CD
MoxiProtect (opens in new tab)
AI content moderation for public figures, founded and built solo
A platform that protects public figures from targeted online abuse, detecting and moderating harmful content across their social channels in real time.
- Real-time harmful-content detection with Gemini, classifying social-media content at scale
- Full stack: Vue.js front end, .NET Core back end, GCP infrastructure, and an Apigee API gateway
- Built hand in hand with the people it protects, iterating on detection and moderation with end users directly
UpgradeMyNight (opens in new tab)
Hospitality and nightlife platform
A platform connecting venues and guests to book and upgrade nights out. Built on Nuxt and deployed on Cloudflare.