Notes from the build.
Engineering deep dives, architecture decisions and what we learned shipping operations-heavy software.
A Plain English Guide to FX Spreads
Your bank quoted a worse rate than the one you saw online, and nobody lied. Here is how spreads work, where businesses lose money on them, and how to measure what you really paid.
A Practical Reading of Nigeria's Data Protection Act for Startups
A plain reading of the Nigeria Data Protection Act 2023 for early stage teams: the ideas that matter, a month one checklist, and where to take proper legal advice.
Small Models, Big Wins
A small model trained on your own data can beat a large one on speed, cost and predictability. When to choose one, and when not to.
Cash Visibility Is the First Treasury Problem
Before forecasting or optimising, a treasury team has to see its cash. Why the picture is blurry, and a practical way to build a trustworthy view.
What Instant Payments Change for Corporate Treasury
Instant rails remove the float, the daily cut off and the window to catch mistakes. What corporate treasury teams should change in forecasting, controls and visibility.
Human in the Loop Is a Design, Not a Disclaimer
Saying a person reviews everything is not a control. Reviewers tire, queues win, and approvals become clicks. What it takes to design human oversight that works.
Reconciliation Is Where Fintechs Quietly Fail
Nothing crashes when reconciliation fails. The gap just grows until someone asks. Why payment companies lose control of their books here, and how to build matching in from day one.
The Super App Question in African Markets
Investors keep asking when you will become a super app. The Asian model rests on conditions that do not always hold in African markets, and other routes can reach the same value.
Structured Output Changed How We Build With LLMs
Moving from regexes over free text to schemas changed how we treat models: as typed components we can test, validate and replace. Here is what it fixed and what it does not.
Why AI Pilots Stall at the Handoff
The demo worked, the model was accurate, and the pilot still never shipped. The failure usually sits where AI output has to become someone's real work.
Patient Records and the Principle of Least Privilege
A receptionist should not be able to open every record at every branch. How to design hospital access around role, need and relationship, without slowing clinicians down.
The Case for Boring Databases
New databases arrive with better stories, but most teams should start with PostgreSQL. A case for boring technology, and the signs that it is time to move on.
Have a problem worth solving?
Tell us what is not working, what you are trying to build, or where you want to go next.