Doctor of Business Administration — Emerging Technology · Concentration in Generative AI
Innovation leader at Hong Kong Customs, award-winning composer and conductor, and Doctor of Business
Administration in Generative AI — one practitioner who has spent more than two decades inside
all three rooms simultaneously.
Composer, public-sector leader, doctoral researcher — the integration claim needs all three.
I write neither as a technologist explaining music to musicians, nor as a musician explaining
technology to technologists — but as a practitioner inside all three rooms at once.
0Years in public-sector tech
0Years as a CASH composer
0Awards & honours
0Professional worlds
🏛️
1. Public-Sector Practitioner
— ✦ —
👤Roles
Head of Project Planning & Development Division II and inaugural Head of the Innovation
and Technology Strategic Development Division (2022–24), Hong Kong Customs —
25+ years in risk management and emerging technologies across Hong Kong's public infrastructure.
🏆Recognition
WCO Disruptive Technology Workshop — Blockchain session moderator (2021)
WCO Data Model Project Team member (2018–2024)
Judging panel, Most Innovative Knowledge Enterprise (MIKE) Award — with HK PolyU's BAKE Research Centre
Keynote speaker & panel member, Asian Knowledge and Innovation Forum (AKIF)
Speaker & coordinator, InnoEX Smart Hong Kong Pavilion (2023 & 2024)
Bronze Winner, Hong Kong ICT Award 2022 — Smart Business (Business & Public Sector Enterprise)
WITSA Global Innovation & Tech Excellence Award — 2023 nominee for Hong Kong, China
Commendation for IPR enforcement with global collaboration
🎵
2. Professional Music Creator
— ✦ —
👤Roles
Composer and author — creator, orchestrator, storyteller — in traditional and
AI-assisted music composition.
🏆Recognition
Member, Composers and Authors Society of Hong Kong (CASH) since 1999 — CASH Music Fund
Former Music Director, TVB — programme production & music composition
Professional violinist, Hong Kong Sinfonietta
Winner, Hong Kong 1st MusicalCon Composition Competition 2026
Bach International Music Competition — Winner in Conducting, 2026
Fenix International Music Competition — Winner in Conducting, 2026
📖
3. Doctoral / AI Researcher
— ✦ —
👤Roles
Doctor of Business Administration — AI researcher, analyst, scholar — trained within
the San Francisco / Silicon Valley generative-AI ecosystem.
🏆Recognition
Doctor of Business Administration in Emerging Technologies, specialised in Generative AI — Golden Gate University, San Francisco (Cohort 5) · graduated 2026
Scholar, MIT Sloan School — AI Executive Academy, Cambridge, Massachusetts
Author, Joint International DBA Studies (JIDS) Journal — EBU European Business Institute, Luxembourg
Founding Council Member & Head of Advisory Board (Emerging Technology), ETHOS Global Charitable Foundation
Active participant, Hong Kong Association of Digital Humanities
✦ Executive Education ✦
MIT Sloan — AI Executive Academy
✦
Scholar of the AI Executive Academy — MIT Sloan School of Management & MIT Schwarzman
College of Computing, Cambridge, Massachusetts — the world's No. 1 MBA, embedded with AI leadership.
Reflections from the Academy
A corner I keep writing in
2026 · Cambridge, MA
“Most people think AI leadership is about learning tools.
I've learned it's actually about learning judgment.”
Completed the MIT AI Executive Academy and earned my Executive Certification in
Digital Business — a journey that was never about collecting credentials, but about
reshaping how we think, decide and lead in an AI-redefined era.
Founding Council Member · Head of Advisory Board, Emerging Technology
ETHOS Global Charitable Foundation
Founding Council Member of ETHOS — a global charitable foundation devoted to
Emerging Technology for Human Optimization & Sustainability —
and Head of its Advisory Board for Emerging Technology, bringing public-sector,
creative and research experience to technology that serves people.
Emerging
Technology
Human
Optimization &
Sustainability
II.The Journey
Two decades in three rooms
✦
Public service and technology on one side; music and scholarship on the other.
Public Service & Technology Pillar I
2011 — 2012
COMPASS & international risk management
Designed COMPASS — the COMputerized Profiling Analysis and Screening System — for risk
management at Hong Kong International Airport, and joined the International Workshop for
Accreditation of Risk Management Advisors (2012).
Systems Design
2018 — 2024
World Customs Organization
WCO Data Model Project Team member and data steward for the Customs Data Model;
moderator of the Blockchain session at the WCO Disruptive Technology Workshop (2021).
International
2021 — 2022
COVID-19 emergency platforms
Designed and led the Community Isolation Facilities Collaboration and Operation (Co-Op
營健通) Platform — a contactless crowd-management cloud platform built with privacy and
ethics at its core.
Cloud & Ethics
2022 — 2024
Innovation leadership & recognition
Inaugural Head of the Innovation and Technology Strategic Development Division, Hong Kong
Customs. Bronze Winner at the Hong Kong ICT Awards 2022; WITSA 2023 nominee;
Speaker & Coordinator at InnoEX 2023–24.
Leadership
Music & Scholarship Pillars II · III
1999
A professional musical life begins
Composer member of CASH; professional violinist with the Hong Kong Sinfonietta and
former Music Director at TVB — Hong Kong's dominant broadcaster.
Music
2026 · Doctorate
Doctor of Business Administration — Emerging Technologies, Generative AI
Graduated from Golden Gate University, San Francisco (Cohort 5), with doctoral research on
generative AI within the Silicon Valley ecosystem; Scholar of the MIT Sloan School
AI Executive Academy.
Graduated 2026
2025
Journal publication & ETHOS founding
Published in the Joint International DBA Studies (JIDS) Journal. Founding Council Member
and Head of Advisory Board (Emerging Technology) of the ETHOS Global Charitable Foundation.
Research
2026
An international year in music
Winner of the Hong Kong 1st MusicalCon Composition Competition and of the Bach and Fenix
International Music Competitions in Conducting; advisory board member of the In-heritage
Philharmonic Orchestra. Writing The Human-Machine Duet.
Milestones
2011 · Devised
COMPASS
COMputerized Profiling Analysis and Screening System — risk management for Hong Kong
International Airport.
Speech recognition, NLP, chatbots at HKIA and boundary control points, video analytics and a
retrieval-augmented generation knowledge base.
III.Recognition
Awards & honours
✦
Milestones across technology and music that I'm grateful for.
🏆
2022
Hong Kong ICT Award — Bronze Winner, Smart Business
Solution for Business & Public Sector Enterprise — the COVID-19 Patient and Caretaker Management Platform.
🥇
2026
Hong Kong 1st MusicalCon Composition Competition — Winner
Winning original composition, with broad local media coverage.
🎼
2026
Bach International Music Competition — Winner, Conducting
International recognition in orchestral conducting.
🎻
2026
Fenix International Music Competition — Winner, Conducting
International recognition in orchestral conducting.
🌍
2023
WITSA Global Innovation & Tech Excellence Award — Nominee
World Innovation, Technology and Services Alliance — nominee for Hong Kong, China.
🎖️
Service
Commendation for IPR Enforcement
Intellectual property rights enforcement with global collaboration at Hong Kong Customs.
IV.On Stage
Talks & appearances
✦
A running diary of my talks and appearances across AI summits.
AI Summit Diary · Featured
Schooling in the AI Era — HKBUAS 20th Anniversary AI Symposium
MC & Facilitator · Panel: “Redefining Educational Leadership in the AI Era” ·
March 2026 · Hong Kong · Featured on China Daily Hong Kong
On stage not just as a speaker, but as MC and facilitator — alongside distinguished
professors and global experts, exploring a shift far deeper than technology adoption:
from AI to Augmented Intelligence.
2026 · Mar
HKBUAS 20th Anniversary AI Symposium — MC & Facilitator
“Schooling in the AI Era — Leading Transformation & Driving Development” · Hong Kong · featured on China Daily HK
MC · Facilitator
2021
WCO Disruptive Technology Workshop — Blockchain Session Moderator
World Customs Organization · guiding transnational discussions on blockchain applications and workflows
Moderator
2023 · 2024
InnoEX — Speaker & Coordinator, Smart Hong Kong Pavilion
International Innovation Expo · Smart Government + Smart Economy + Smart People solutions
Speaker
Recurring
Asian Knowledge and Innovation Forum (AKIF)
Keynote speaker and panel member · knowledge management and innovation across Asia
Keynote
Recurring
Most Innovative Knowledge Enterprise (MIKE) Award — Judging Panel
With the Hong Kong Polytechnic University's Behaviour and Knowledge Engineering (BAKE) Research Centre
Judging Panel
2012
International Workshop for Accreditation of Risk Management Advisors
International accreditation workshop · risk management profession
Workshop
Industry
AI-Assisted Music Production Workshops — Producer & Performer
Professional workshops bridging generative AI and live music creation
Workshop
V.Research Output
Journal & publications
✦
Peer-reviewed research and writing in progress.
Can Transformers Transform WCO Data Model Interoperability? A Transformer-Based, Human-in-the-Loop Pipeline: A Design-Science Ex-Ante Evaluation
Kwong, J. — Joint International DBA Studies (JIDS) Journal, Vol. 1, No. 002 (2025) · EBU European Business Institute, Luxembourg
A transformer-assisted mapping pipeline that resolves semantic inconsistencies and version
compatibility in the WCO Data Model — combining automated retrieval and ranking with mandatory
validation constraints and human oversight, evaluated through design-science methodology for
safe AI deployment in the public sector.
📖 The Human-Machine Duet: The ABCDE Guide for GenAI Creativity
Kwong, J. — Book in progress, anchored in doctoral research (Golden Gate University, San Francisco)
A practitioner's guide to generative AI creativity across five layers — Artificial &
Augmented Intelligence, Blockchain & Big Data, Cloud Computing, Data Strategy and Ethical
Governance — treating provenance, value exchange and governance as one integrated
infrastructural problem, validated in the music ecosystem.
Coming Soon
VI.Visual Library
AI & concept diagrams
✦
Diagrams and infographics from my LinkedIn — click any of them to open the full post.
Attention Residuals
Rethinking the transformer's most fundamental assumption — click to read the post.
AI Assistance & Human Persistence
What AI help gives today — and can quietly take away tomorrow.
Impact Planet for Generative AI
A planetary radar of maturity, momentum and mass — Human Optimization at the core.
Creativity Connects
The ABCDE of GenAI creativity — technology, creativity, humanity.
VII.Writing
Blog & insights
✦
My writing on AI — from LinkedIn to here. Click a post to read it in full.
LLM Architecture · LinkedIn
Attention Residuals: rethinking the transformer's core assumption
Transformers once used attention to decide what words matter. Now they're learning to decide what thoughts matter across layers.
Golden Gate UniversityMIT Sloan — AI Executive AcademyETHOS Global Charitable FoundationWorld Customs OrganizationEBU LuxembourgAKIFHK PolyU · BAKEInnoEXWITSACASHIn-heritage PhilharmonicHong Kong SinfoniettaTVBCUHK
Let's compose the AI future together
✦ ✦ ✦
Invite me to speak, collaborate on research, or explore AI, innovation and creativity for your organisation.
The recent paper “𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻 𝗥𝗲𝘀𝗶𝗱𝘂𝗮𝗹𝘀” introduces a rethink of residual connections — the backbone of every modern transformer.
A subtle but important shift in LLM architecture just dropped — and it challenges one of the most fundamental assumptions in deep learning.
The problem: today's models blindly accumulate outputs from every layer with equal weight. Over depth, this creates “signal dilution” — earlier insights get buried, and later layers must overcompensate.
The breakthrough: replace static residuals with attention over layers. Each layer dynamically decides which prior representations matter, using learned, input-dependent weights.
In short: 👉 Models stop treating all past computation equally 👉 Depth becomes selective, not cumulative 👉 Information retrieval across layers becomes intentional
To make this practical at scale, the authors introduce Block Attention Residuals, reducing memory overhead while preserving gains — making it viable for production-scale LLMs.
Why this matters for AI leaders: we're entering the era of “depth efficiency”. Scaling is no longer just about more layers — it's about using depth intelligently. Architectural innovation is back at center stage: after years of data + compute scaling, we're seeing meaningful gains from rethinking core primitives. And there are implications for cost-performance curves — if models can extract more value per layer, this could materially shift training and inference economics.
A broader pattern: attention is expanding beyond tokens → now across time, modalities… and depth itself.
Transformers once used attention to decide what words matter. Now they're learning to decide what thoughts matter across layers.
That's a meaningful step toward more adaptive — and ultimately more efficient — intelligence.
AI assistance may improve performance today — while weakening independent performance tomorrow.
A recent preprint by Grace Liu, Brian Christian, Tsvetomira Dumbalska, Michiel A. Bakker and Rachit Dubey raises an important question for the AI era: what happens when a system is optimized to give immediate help, but not necessarily to build long-term human capability?
Across three randomized controlled trials with 1,222 recruited participants, the researchers found a consistent pattern. AI assistance improved short-term performance during the assisted phase. But when AI was removed, participants performed worse on their own and were more likely to give up. The effect appeared after only about 10 minutes of AI-assisted interaction.
The paper's central insight is not that AI is harmful by nature. It is that current AI assistants may act as short-term collaborators: highly useful in the moment, but not necessarily designed to protect the user's future autonomy, persistence, or independent reasoning.
This distinction matters. A good mentor does not simply provide answers. A good mentor scaffolds learning, calibrates support, tracks progress, and sometimes withholds direct help so that the learner can develop the capacity to persist. The concern raised by the study is that AI systems, when used mainly as answer machines, may remove the very struggle through which skill and confidence are built.
One finding is especially important: the downstream decline was concentrated among participants who used AI to obtain direct solutions. Those who asked for answers showed lower later performance and lower persistence than those who used AI for hints or understanding.
The risk is not AI assistance itself. The risk is unscaffolded assistance.
For education, work, and organizational learning, the implication is clear. We should not judge AI only by immediate task completion. We should also ask whether it strengthens the human capacity to perform when the tool is absent.
The better design question is not “How fast can AI give the answer?” — it is “How can AI help people remain capable, persistent, and independent after the answer is gone?”
That may be one of the most important design challenges for responsible AI: building systems that do not merely complete tasks for humans, but help humans preserve the ability to think, struggle, and learn.
Source: Liu, G., Christian, B., Dumbalska, T., Bakker, M. A., & Dubey, R. (2026). AI Assistance Reduces Persistence and Hurts Independent Performance. arXiv:2604.04721.
When I first looked at the Generative AI impact radar (from Gartner), I did not only see a technology landscape. I saw a planet. And the centre of this planet should not be the model. It should be Human Optimization.
Each of the dots has its own maturity timeline. Each has its own weight of impact. Each may reshape how organizations think, build, decide, and serve. The real question is not simply: which technology will mature first? The deeper question is: which will truly augment human capability, improve institutional intelligence, and create responsible value?
That is why I placed Human Optimization at the core, and Human Augmentation on the crust. Because AI should not only automate tasks — it should extend human judgment, creativity, productivity, and decision-making capacity.
At the same time, this planet is not calm. There are earthquake shockwaves: employment and organizational changes. There are volcanoes: eruptions of new models and game-changing use cases. There are vast oceans: a vast ocean for use cases and markets, for research, and an even deeper ocean for responsible governance.
And a warning from the build-and-data side: Garbage In, Garbage Out. Data quality, governance, provenance, and contextual grounding are no longer back-office technical issues. Poor data does not only create poor outputs; it creates poor decisions, poor trust, and poor institutional learning.
For leaders, the real leadership challenge is to navigate this planet wisely: to identify what is ready now, what is emerging next, what requires governance before scale, what may disrupt people and structures, and what can genuinely create human-centered transformation.
GenAI will not be won by those who only chase the newest model. It will be shaped by those who can connect models, data, governance, workflow, human judgment, and organizational purpose into one coherent operating system.
The destination: where technology does not replace the human mission, but expands what humans and organizations can responsibly achieve.
That's why I relearn and create the value for our future generations, through my own study and creation of the global collaboration ahead.
#HumanOptimization #FutureOfWork
Artificial Intelligence and Augmented Intelligence, Blockchain & Big Data, Cloud Computing, Data Strategy, Ethical Governance.
Technology. Creativity. Humanity. Building a harmonious future.
The concept behind my doctoral research and my book in progress, The Human-Machine Duet: The ABCDE Guide for GenAI Creativity — where the five layers play together like a chord, and creativity connects them all.
Honoured to have taken the stage — not just as a speaker, but as the MC and facilitator — alongside distinguished professors and global experts at this meaningful gathering on Educational Innovation in the AI Era.
Together, we explored a shift that is far deeper than technology adoption. It is a shift in how we think about intelligence itself.
At the heart of my sharing — “From AI to Augmented Intelligence” — is a simple but urgent proposition:
We are not here to replace human intelligence. We are here to elevate it.
In a world increasingly shaped by generative AI, the real question is no longer: “What can AI do?” But rather: “What should humans become — with AI?”
Across the dialogue, several themes strongly resonated:
🔹 From knowledge transfer to knowledge transformation. Learning is no longer about consuming information, but about synthesising, contextualising, and applying it with judgment.
🔹 From individual intelligence to collective intelligence. With AI as a collaborator, classrooms, organisations, and societies are becoming ecosystems of co-creation.
🔹 From tools to trust. The future of AI in education is not defined by capability alone — but by governance, ethics, and responsibility.
As MC and facilitator, what struck me most was not just the diversity of perspectives — but the convergence of conviction:
👉 The future of education is not AI-driven. 👉 It is human-centered, AI-augmented.
This aligns deeply with my ongoing doctoral journey and practice: AI is not the destination. Augmented Intelligence is.
We are not just building smarter systems — we are shaping smarter humans, smarter institutions, and a more thoughtful society.
Grateful to be part of this dialogue — and even more excited about what comes next. Let's continue to build, question, and co-create.
Most people think AI leadership is about learning tools. I've learned it's actually about learning judgment.
Honoured to have completed the MIT AI Executive Academy and earned my Executive Certification in Digital Business. This journey was never about collecting credentials. It was about reshaping how we think, decide, and lead in an AI-redefined era.
What makes this experience truly distinctive is the intellectual depth behind it. The Academy is delivered by esteemed professors and scholars across a wide spectrum of academic bodies within the Massachusetts Institute of Technology (not exhaustive), including:
🏛️ Schools & Colleges
MIT Sloan School of Management
MIT Schwarzman College of Computing
MIT School of Engineering
MIT School of Humanities, Arts, and Social Sciences
MIT School of Science
🔬 Departments & Academic Disciplines
Electrical Engineering and Computer Science
Management
Aeronautics and Astronautics
Economics
Operations Research and Statistics
Information Technology
Marketing
Applied Economics
🧠 Research Labs & Institutes
MIT Computer Science and Artificial Intelligence Laboratory
MIT Laboratory for Information and Decision Systems
MIT Institute for Data, Systems, and Society
MIT Jameel Clinic
MIT Generative AI Impact Consortium
MIT Center for Quantum Engineering
Research Laboratory of Electronics
What I take away — in this rocket-paced AI era, one conviction has become even clearer: we should not stop at using AI. We must learn to govern it, contextualize it, and augment with it.
From Artificial Intelligence → Augmented Intelligence: Elevate human judgment, Anchor responsibility, Build systems of trust.
The real value — beyond the curriculum, what made this journey exceptional was the people. Deeply grateful for the wisdom and truly practical insights shared by an incredible global cohort spanning 19 countries. It felt less like a classroom, and more like a living system of global intelligence.
In the age of AI, the real competitive advantage is not intelligence — it is fast, responsive and responsible judgment.
Moving forward — the question is no longer “What can AI do?” but “What should we do with AI — responsibly, ethically, and at scale?” This milestone continues to shape my journey in bridging AI, governance, and human-centered design — into what I call The Human–Machine Duet.
