What a pitch competition actually tests
After QuantumByte won SelectUSA Indonesia 2026: what judges are really probing beneath the slides, and why it is the same thing a serious client checks.
Read What a pitch competition actually tests · 3 minWriting
Notes on how I work, written while doing it.
Short, practical posts from the projects: how discovery runs, how a scope is written, how AI stays inside your rules, and what a team actually pays for. Newest first. Subscribe by RSS.
The numbers I will and won't put on a slide
After the AIIP white-paper launch: the difference between reported and measured outcomes, how to read a case study, and why this site labels where every figure comes from.
Read The numbers I will and won't put on a slide · 3 minSmall observable slices
After live product demos at the Tech in Asia conference: how enterprise software gets shipped without a big-bang launch, one beneficiary and one outcome at a time.
Read Small observable slices · 3 minKeeping AI inside the lines
Catalog-only outputs, confidence and reasoning beside every match, human review, and batch visibility: the pattern that makes AI usable inside a large organization.
Read Keeping AI inside the lines · 3 minTeenagers shipped real apps in two days
What QuantumByte's two-day workshops with high-school students and village participants proved about who should build operational tools, and what the tools have to be like.
Read Teenagers shipped real apps in two days · 3 minAcceptance is not adoption
Acceptance, adoption, value, and renewal are four different events. Confusing them is how software projects are declared successful and quietly abandoned.
Read Acceptance is not adoption · 3 minFrom data to decisions
After a talk on machine learning from data to decisions: why a recommendation is only useful with its reasoning beside it, and how leaders actually act on briefs.
Read From data to decisions · 3 minWhat a national platform taught me about scope
From the INA Digital launch-planning discussions: why decision rights, acceptance authority, and the smallest end-to-end outcome matter more than the feature list.
Read What a national platform taught me about scope · 3 minDesign the operator's day before the database
Why I map a dispatcher's or a driver's day from trigger to end state before touching the data model, and what changes when you do.
Read Design the operator's day before the database · 3 minJourno: a pipeline that turns articles into places
How I built Journo for Postcard: per-source crawlers, parsers, spaCy enrichment, SerpApi and Who's On First matching, a scoring threshold, and an operator dashboard.
Read Journo: a pipeline that turns articles into places · 3 minOne year in: what enterprise clients actually pay for
After a year of building software for large organizations, the three things a buyer is really checking, and none of them is the feature list.
Read One year in: what enterprise clients actually pay for · 3 minShow me the last time this happened
Why I stopped asking teams to describe their process and started asking for the most recent real case, with the six questions that follow.
Read Show me the last time this happened · 3 minThe rule I set on day one: revenue before headcount
Two months into founding Quantum with $20K of my own money, the one rule that decides who gets hired, what gets built, and which projects I turn down.
Read The rule I set on day one: revenue before headcount · 3 min
Bring me the problem. I’ll bring the plan.
Thirty minutes. Bring the last time it went wrong. You leave with a straight answer.