I build secure cloud infrastructure, edge applications and AI-powered tooling — and help organisations design, adopt and operate them, from the first architecture conversation to production.
I believe secure, high-performance applications are the result of deliberate design and disciplined engineering.
I work at the intersection of hands-on engineering, architecture and customer success. I don't just build secure cloud-native systems, edge applications and AI-powered tooling — I help teams design, adopt and operate them well, with a focus on security, simplicity and operational excellence.
That means I'm as comfortable hardening an edge deployment or wiring up an agentic AI workflow as I am running an architecture review, scoping a proof-of-concept or turning a business goal into a pragmatic technical plan. Twenty years across telecom and enterprise infrastructure taught me to understand the customer's problem first, and choose the technology second.
Each project, from the problem it solves to how it's built and what it delivers.
Built a production SaaS platform that automates UK parking appeals.
UK drivers get confusing parking charge notices and rarely appeal, even when they'd win.
AI SaaS on Cloudflare Workers — LLM-generated appeal letters, with Cloudflare D1 (SQLite) and R2, protected by a WAF with managed and custom rules.
A photo of a notice becomes a complete, ready-to-send appeal letter in minutes.
Reduced Kubernetes troubleshooting time from hours to minutes with an AI-powered diagnostic agent.
Diagnosing failing pods in a k3s cluster is slow, manual and token-hungry for AI agents.
A declarative AI agent (kagent + MCP) with health-check, diagnose and heal modes — LLM-agnostic, defined in a single YAML.
One-command diagnosis with 75% fewer tokens (46K → 12K). Shared at MCP_HACK//26.
Runs a live, self-hosted platform serving production-like and personal workloads around the clock.
I needed a production-like environment to practise cloud-native and edge skills with real workloads.
A 3-node K3s HA cluster plus Proxmox virtualisation and Zero Trust ingress, self-hosted on bare metal.
An always-on lab running live services, published at igorbond.info.
Prototyped an edge API that securely collects raw e-bike telemetry and, within a hundred milliseconds, sends AI-controlled commands back to the e-bike.
E-bike telemetry and riding-session data must be processed as quickly as possible (within hundreds of milliseconds), and deterministic or AI-driven command chains should feed back within a few seconds.
An edge API pairing a fast deterministic decision loop with an agentic AI loop, backed by persistent edge storage, issuing commands back to the bike.
The main objective was to evaluate whether hundred-millisecond performance for a deterministic decision loop, and a few-second performance for an agentic AI inference loop, could be achieved.
How I work with customers and teams — whether I'm building it myself or helping them build it.
Understand the business goals, technical constraints and the architecture already in place.
Shape secure, scalable solutions with pragmatic trade-offs made in the open.
Prototype, automate and deliver production-ready infrastructure.
Document, teach and support the team so they can run and evolve it long-term.
The customer's goal and constraints come first; the tools are chosen to fit them, not the other way round.
Security is a starting assumption in every design, not a layer bolted on at the end.
The best architecture is the simplest one that meets the need and is easy to reason about under pressure.
If I can't see it, I can't run it — so metrics, logs and traces come first, not last.
Automate the repetitive and the risky where it genuinely saves time or removes error — never for its own sake.
A solution only succeeds if the team can run and evolve it after I've moved on.
I use AI to move faster and think wider, while owning every decision it informs.
And I'm always learning and adopting new ones.
Virtualised and bare-metal infrastructure deployment for multi-site 4G/5G Core functions — and technical interface to telecom/enterprise customers.
Service manager for enterprise networking and telecom infrastructure across the holding's operating companies.
Group-wide network architecture and strategy alignment across multiple operating companies and key vendors.
17 years' progression across core engineering, service management and platform architecture.
Open to interesting projects — on site or remote. If you're building something where technical depth meets customer-facing delivery, let's talk.