Perspectives from practice
I write about technology leadership, AI architecture and the decisions behind the systems I build.
The focus is practical: what agents should be allowed to do, where people remain accountable, and how we know the systems we rely on actually work.
These are the full pieces. Shorter versions published elsewhere point back here as the version of record.
Agents need direction, not motivation
Leading people and directing agents are different disciplines. What changes, and what remains a leadership responsibility.
One proof isn’t enough
Why observing an agent isn’t enough, and how we prove that what reaches production actually works.
Memory agents can read, but not quietly rewrite
Why useful memory needs a boundary between what an agent can know and what it is allowed to change.
Reading Singapore's agentic AI framework as an architect
The parts of Singapore’s agentic AI framework that change how I think about architecture, authority and accountability.
What working with agents taught me about leadership
What agent-heavy engineering taught me about clarity, delegated authority and the judgment that still belongs to people.
Someone always had to sign
What Singapore’s interbank settlement system taught me about approval, evidence and decisions that have to remain explainable years later.