This site is currently under major renovation
Teaching philosophy
I want to build things that other people actually use, and I teach from that test: was it adopted, and did it add value to the system it went into?
The first computer I used was an Apple II. I have been interested in computing ever since, and the interest has always been of a particular kind: I want to build things that other people actually use. That is still the test I apply to my own work. A tool nobody uses has taught its builder very little. I teach from that position, and everything below follows from it.
1. Practical work combined with applied learning
My teaching rests on the combination of practical work and applied learning. The two are not the same thing. Practical work means students build. Applied learning means what they build is answerable to a real context, with real constraints and someone real on the other side of it. Theory is introduced at the point where the build demands it, because that is the point at which a student can actually hold it. I have taught this way in a vocational institute, a design institute, a transnational degree programme, an executive MBA and a corporate training room, and the method transfers across all of them.
2. Depth or breadth, chosen deliberately
Work succeeds either by going deep or by going wide. Both are legitimate; what is not legitimate is doing neither by accident. My own record shows both moves. On the depth side: a doctorate built on a single empirical question, and years spent taking one clinical system into more than 200 dental clinics until it worked in every one of them. On the breadth side: computing, law, finance, education and enterprise delivery, held together because each one explains something the others cannot. What I try to teach students is the judgement to tell which move a piece of work calls for, because most weak student projects are shallow in both directions at once.
3. All-round graduates, rooted in one discipline
I want graduates who are all-round but rooted. The root should be one professional discipline held to a professional standard, computer science in this case. From that root a graduate can move outward in many directions: advisory work, consultancy, system development, the analytical discipline a lawyer applies to evidence, financial and risk planning. I hold degrees in law and in financial management alongside computer science, and I did not acquire them to leave computing. I acquired them because a computing professional who can read a contract, price a risk and argue from evidence is a more useful person to have in the room. The discipline is a root system, not a ceiling, and I say so to students explicitly, because many of them assume the degree defines the ceiling.
4. Fundamentals taught through consequence
Operating systems, data structures, algorithms and systems programming are not topics students should merely pass. They are the layer at which everything above them eventually fails. When I taught operating systems and network administration in Hong Kong, I framed each abstraction against a failure I had actually seen in the field: a scheduler starving a process in a live clinical deployment, a subnet misconfiguration taking a pilot system offline. Students remember the concept because they remember the consequence. Complexity analysis is not an exam ritual; it is the reason a query that ran on 200 test records collapses at 200,000.
5. Real clients, not realistic ones
The strongest learning I have observed comes from work with a genuine external stakeholder. Redesigning CSC2102 Professional Software Development and Team Project at the University of Glasgow (Singapore), I built a simulated industry model around real partners: 60 industry-linked projects with 28 partner organisations, about 150 students per cohort, an 86% completion rate, and work-based learning grown from 90 to 160 students per cohort. Students negotiate requirements with someone who actually wants the software, hold reviews with them, and ship. The difference between a real client and a client played by a tutor is the difference between learning professional judgement and learning to satisfy a rubric.
6. Feedback is the mechanism, not the paperwork
Feedback that arrives after the work is finished is documentation, not teaching. This is a scale problem, so I treated it as an engineering one. I designed and deployed an AI-assisted peer-evaluation platform now used by 300+ students a year, processing 3,500+ feedback entries per trimester. It cut staff marking turnaround from about two weeks to three days and the student feedback cycle from about two months to two weeks, and it has been adopted by five universities. I also introduced a Dynamic Examination System generating individualised but equivalent assessments over three academic years, which protects integrity without resorting to surveillance. My position on generative AI is the same: used transparently and with guardrails it makes feedback faster and fairer; used to replace the judgement of the person teaching, it degrades both.
7. Teaching the whole range of learners
I have taught vocational and polytechnic students, undergraduates, postgraduates, evening adult learners and corporate professionals. Those audiences need genuinely different things. A vocational student needs to see the machine before the theory. A mid-career professional converting into technology needs the theory compressed and the relevance made explicit in the first ten minutes, because their tolerance for abstraction is low and their standard for usefulness is high. Designing for that range, including a digital badge and micro-credential initiative that issued more than 2,000 badges across the Hong Kong Vocational Training Council, has made me a better undergraduate teacher, because it forces you to know why each thing is being taught rather than merely that it is on the syllabus.
8. Curriculum should move when industry moves
I introduced Rust into the Glasgow Singapore curriculum because the systems and security-skills demand was visible in industry well before it reached textbooks, and I ran a controlled study of an IDE plugin for novice Rust programmers so that the introduction was evidenced rather than fashionable. As an ABET and CSAB Programme Evaluator, working with around ten institutions across eight countries, and as External Examiner at THEi Hong Kong, I spend part of each year examining how other departments build and defend a curriculum. It is the most useful professional development I have had. Refreshing a curriculum is not chasing novelty: each addition should displace something that has stopped earning its place and should be evaluated afterwards.
9. Advising, and holding myself to the same standard
Much of the value I add happens outside timetabled hours: supervising final-year and capstone projects, several of which have become published papers with the student as first author, and running competitions that give students a reason to build beyond assessment. I founded the Glasgow Singapore hackathon and scaled it from 40 participants across three polytechnics to 90 across all five. Having founded and run a software company for a decade, I can also tell a student honestly what starting something actually costs. I ask students to evidence their claims, so I evidence mine: Senior Fellowship of Advance HE, a Master of Education in Academic Practice, workshops for colleagues on the practical use of AI in teaching, and published work on my own teaching. Teaching that is not studied is teaching that cannot improve.
In summary
I want students who can reason about a system from the silicon up, who have built something a real user adopted before they graduate, who have received honest feedback quickly enough to act on it, and who understand that their discipline is a root rather than a boundary.
Evidence for each claim: Academic work and teaching, Design Portfolio and Experience. Related: Research statement.
