Background

30+ Years in IT

The career started before commercial internet access was a thing — running BBS systems, managing small communities across dial-up lines, and learning that networks break in the most inconvenient ways possible. Those habits of thinking about failure modes and edge cases have never left.

Across three decades the work has spanned systems engineering, network architecture, large-scale automation, cloud platform engineering, and now AI systems integration. The common thread is complex systems with real consequences when they stop working.

The current engagement with AI is not a career pivot. It is a continuation of the same work — understanding how new technology actually performs under production conditions, what the real limitations are, and where the engineering effort belongs.

There are no vendor relationships here. No advisory board seats. No consulting agreements that create conflicts of interest. What is written is an honest assessment of what the technology does and does not do.

30+
Years in IT
Continuous hands-on experience
~1993
Career start
BBS systems, pre-commercial internet
AI/ML
Current focus
LLMs, inference, automation, RAG
None
Affiliations
No vendors, advisors, or sponsors

Philosophy

How This Site Approaches AI

Real-World Usage

Theory without production experience is speculation. Everything written here is grounded in what actually works when you move past the demo and hit real scale, real data, and real failure modes.

Cutting Through Hype

The AI space is saturated with marketing dressed as analysis. Benchmarks are curated. Press releases omit limitations. Independent assessment requires no loyalty to vendors or their narratives.

Practical Outcomes

The question is never just "can it do this?" but "should we do this, at what cost, and what breaks when it fails?" Engineering is about trade-offs, not capabilities in isolation.

Experience

Technology Timeline

Early 1990s

BBS Sysop

Running bulletin board systems. Dial-up networking, FIDONET, early peer-to-peer file sharing, and community management before the public internet.

Mid–Late 1990s

Systems & Network Engineering

Transitioning from BBS to TCP/IP networks. Windows NT, Novell, early Linux deployments. First exposure to enterprise-scale infrastructure.

2000s

Enterprise IT & Automation

Infrastructure architecture, scripting and automation, data centre operations. Building systems that had to survive years of production load.

2010s

Cloud & Platform Engineering

Cloud migration, DevOps, containerisation, CI/CD pipelines. The shift from owning hardware to orchestrating services.

2020s – Present

AI Systems & LLMs

Hands-on work with large language models, inference infrastructure, RAG systems, AI automation, and agent orchestration in production environments.

Mission

Why Does This Site Exist?

The AI space has a signal-to-noise problem. Most of what is published is either vendor marketing, breathless speculation, or academic work that never touches production systems. There is very little from engineers who are actually building things and have enough scar tissue to know what matters.

This site is an attempt to contribute to that smaller category. It will not always be right. It will be honest about uncertainty. It will not be comprehensive — there is no ambition to cover everything. But what it does cover will be grounded in real experience and written without an agenda.

If you are an engineer, architect, or decision-maker trying to separate genuine capability from marketing noise, this is for you.