About Me
I build custom voice AI systems. Based in Folkestone, UK.
Most of my work involves migrating companies from managed platforms (Vapi, Retell) to open-source frameworks (Pipecat, LiveKit). Beyond the obvious cost savings, you get data sovereignty, vendor independence, and crucially, the flexibility to adapt as the space evolves. When running models locally becomes viable for your use case, you're already positioned to take advantage of it.
I specialise in:
- Debugging production voice AI systems (CPU leaks, latency issues, resource management)
- Architecting migrations from managed platforms to self-hosted infrastructure
- Building configuration abstraction layers for multi-vendor systems
- Turn-taking optimisation and interrupt handling
- Balancing low latency with non-interruptiveness through careful tuning of VAD, Smart Turn Detector, and STT/LLM/TTS combinations
I work in code: Python for the backend (Pipecat, LiveKit), Next.js for frontends, Postgres (Supabase, Neon) or libSQL (Turso) for data.
I handle DevOps too: I deploy across VPS, serverless (Fly, Railway), managed (Pipecat/LiveKit Cloud), or self-managed infrastructure as needed.
I don't believe in 6-month development cycles. This entire website, voice demo, and its observability was built within 72 hours during a holiday season sprint. I bring that same velocity to client projects.
Check out examples of my work on GitHub, LinkedIn, X and YouTube.
Life Before AI
My background spans a range of technical domains. After earning my Master's in Computing from Imperial College, I built everything from content management systems and trading floor software to mobile apps and smart contracts. This variety taught me to quickly adapt to new technologies and recognise patterns across different systems; skills that prove invaluable in the rapidly evolving voice AI space.
Life Beyond AI
I'm based in Folkestone for the coastal lifestyle. When I'm not working, you'll often find me along the seafront or exploring the local trails.