Kelvin Kaifung Chan

I'm co-founder at Interfacing Research Laboratory, where I build digital twins that make professional practices and SMEs legible to AI, giving agents the operating context they need to support real work.

Prior to software, I trained as an architect, studying planetary systems, remote sensing, simulation, and digital twins as ways of analysing and modelling complex environmental systems.

Some things I have/am currently working on:

Proximity

Operating intelligence for professional practice. A forward-deployed platform that makes your work observable, your decisions grounded, and your actions governed around how your team actually works.

MONOid

Plan the week and review the week in one space across your domains, synced from the tools you already use.

WyrdOS

A calm opinionated operating system for teams who want clarity before velocity: ethos, principles, goals, habits, and projects in one legible stack.

Pilfer

Treat spend as a design decision: capture intent before you buy, and keep scouting, assets, and runway in context instead of losing it to spreadsheets.

IRL PLAYground

Software and toys to help early learning teachers turn play into adaptive, context-specific learning.

Some writing on operating context, agents, systems, and professional judgment:

Context is not hiding in your database

Proximity / Concepts / June 24, 2026

A useful corrective to the idea that context is just stored data.

Why Interfaces Need Operating Boundaries

Proximity / Methodology / June 24, 2026

On where interfaces end and governance begins.

What is operating intelligence?

Proximity / Concepts / April 14, 2026

The clearest doorway into the Proximity thesis.

The difference between automation and delegation

Proximity / Methodology / March 28, 2026

A clean distinction for agents that do work, not just tasks.

Past projects and research: