About

Hi — I’m Nam Chan (陳冠南). Depending on where we met, you might also know me as Alfred Nam. I read Computer Science at The Chinese University of Hong Kong, and I’ve spent the decade since building software — first with my hands, then with teams, and now, increasingly, with agents. I live between London and Hong Kong.

I’m the co-founder and CIO of AceSpeak, and the CTO of Wistkey.

AceSpeak: our dream is getting close

AceSpeak is an AI interview and public-speaking coach. A person uploads a video take and receives a delivery score with specific, encouraging guidance for the next one. The product is built around a simple practice rhythm: see what happened, change one thing, and try again.

The larger dream is even simpler: in the age of AI, help people rediscover — and enjoy — the joy of speaking. AI can help write the words, but a real person still has to walk into the interview, stand up for the presentation, make the pitch, and say them.

From June to August 2026, our team has been working intensely to turn that belief into a product people can actually use. Every test, retake, difficult product decision and honest piece of beta feedback has moved the dream a little closer. We are now preparing for a late-August launch. I am excited, grateful, and very aware that launch day will be the beginning of the work, not the end of it.

Wistkey and production AI

I’m the CTO of Wistkey — a London-headquartered engineering firm that builds and runs software for enterprise and government across the UK and Hong Kong. We’re a partner, not a placement agency: we embed with our clients’ teams and own the outcome alongside them.

Three things sit on my desk:

• Production AI — large language models, autonomous agents and intelligent workflows that run in production, not in a demo. RAG architectures grounded in a client’s own data, fine-tuned models, multi-agent orchestration with tool use, guardrails and human-in-the-loop checkpoints.
• Private, on-premises deployment — LLMs on a company’s own servers, so sensitive data never leaves the building. This is what makes AI possible for the government agencies, banks and healthcare bodies we work with, where data sovereignty isn’t optional.
• Engineering delivery — on-demand engineers and dedicated teams, coordinated from London across delivery hubs in Hong Kong, Shenzhen and South-East Asia.

Some numbers from the work, because they’re the honest measure of it: ten years of enterprise delivery, 120+ clients, 50+ projects shipped, 15+ government and public-sector platforms, and systems now carrying over two million active users.

The work I’m proudest of

Much of it is the kind of public infrastructure you’d never think about until it breaks. Most of it sits under client confidentiality, so I won’t name names — but the shape of the work is consistent: systems in regulated environments, where the data is sensitive, uptime isn’t negotiable, and the security review takes longer than the build.

Unglamorous, heavily governed, and it has to work at 3am. That’s the kind of engineering I like.

AI adoption is a leadership problem, not a software rollout

More and more of my time goes into helping companies adopt AI across the whole organisation — not just the tech department. Finance, HR, admin, accounting, operations: every one of these teams has repetitive, rules-based work that AI can take on, and that’s usually where the fastest wins hide.

The hard part is rarely the model. It’s deciding how work changes, what responsible practice looks like, how managers coach it, and where stronger capability takes someone’s career. Companies that only update the technology stack — and not the management and career design around it — tend to lose their most AI-fluent people.

I teach and speak on AI when people ask — most recently on agentic AI for finance professionals — and I genuinely enjoy it.

Back to code

After some years away from hands-on programming, I’m hooked on it again. AI has re-empowered the skills I built long ago and, honestly, made coding physically more affordable — the long, unbroken hours that get harder to sustain past thirty are no longer the price of shipping something good. That’s changed everything about how much I can build.

So I build in public. I rebuilt my wife’s WordPress site as 102 pages of static HTML in a weekend, took a Google Cloud bill from $2,800 to $300 in an afternoon, and ship small tools like NamLoop between bedtime and sleep. It’s all on the projects page, and I write up how it was done — including the parts that went wrong.

Yoga

I practise Ashtanga yoga, five to six mornings a week, at 4am London time in Euston. It’s the discipline that keeps everything else sustainable — body and mind both. Huge thanks to my teachers Hamish and Clayton for their guidance; the practice I have today is thanks to them. 🙏

What I write about

This blog is where I think out loud about tech, IT, AI, business, finance and yoga — building AceSpeak, practical AI adoption and on-prem LLMs, the craft of shipping software, building companies and making sense of money, and the practice that keeps me grounded. No hype, and I show the working.

I write in English and Traditional Chinese — the 中 / EN toggle in the header switches between them.

Say hi

If you’re wondering what AI could actually do inside your organisation — or what it shouldn’t — that’s my favourite conversation. The fastest way to reach me is email: nam@wistkey.com. Everywhere else I exist on the internet is on the links page.