Search for a command to run...
Series
A series on the architecture decisions that separate a working AI demo from a production system. Drawn from shipping TaxLens (income tax estimation from bank statements), TrustRail (BNPL underwriting without credit bureaus), and a connected set of open-source libraries — each article goes deep on one decision: how to ground language models in domains where being wrong has real consequences, how to control cost before the bill arrives, how to test non-deterministic dependencies without flaky CI, how to structure a frontend that survives feature growth, and how independent applications communicate across team and framework boundaries. Written for engineers who already know what an LLM is and want to understand what production with one actually looks like.
Heyyyyy, Picture this: you're working on a large frontend platform. Multiple teams. Multiple apps. Each team owns their microfrontend slice of the product, deploys it independently, and has the autono
A few months ago I shipped two products within weeks of each other. One computes your Nigerian income tax from a bank statement. The other decides whether a small business gets a BNPL loan. Different
There is a class of test suite that is worse than having no tests at all. It passes most of the time. It fails occasionally with no clear reason. It passes again after a retry. You trust it when it pa
The first time you check your OpenAI bill after a real traffic spike, something changes in you permanently. It's not the number itself it's the realisation that every engineering decision you made in
Most of the world's credit infrastructure is built on an assumption: the borrower has a credit file. A bureau has their repayment history. A score exists. A lender queries it. In Nigeria — and across
Most articles about Feature-Sliced Design show you the theory. The layer diagram. The import direction rules. The terminology (slices, segments, public API). Then they show you a contrived example wit
