Skip to content

migration

strangler fig in practice article

Engineering

The strangler fig pattern, in practice

Incremental migration sounds obvious until you have to route live traffic. Here is what it actually takes.

evaluating llm features article

Data & AI

You cannot ship an LLM feature without an eval suite

Manual spot-checking does not scale past the demo. Offline evals are the difference between a prototype and a product.

slos that mean something article

Cloud & Reliability

SLOs that mean something to the business

Ninety-nine point nine of what, measured where? Most SLOs measure the wrong thing.

design systems that survive article

Design & Product

Design systems that survive the handover

Most design systems decay within a year. The ones that do not share three properties.

postgres is probably enough article

Engineering

Postgres is probably enough

Before adding a fourth datastore, check what Postgres already does.

cloud cost is an architecture problem article

Cloud & Reliability

Cloud cost is an architecture problem

Rightsizing gets you 20%. Architecture gets you the rest.

accessibility is cheaper early article

Design & Product

Accessibility is cheaper when you start early

Retrofitting WCAG compliance costs several times what building it in does.

observability beyond dashboards article

Cloud & Reliability

Observability beyond dashboards

If you only know what you thought to graph, you do not have observability.

retrieval quality beats model choice article

Data & AI

Retrieval quality beats model choice

Teams spend weeks comparing models and days on their retrieval pipeline. The ratio should be reversed.

secure sdlc without slowing down article

Security

A secure SDLC that does not slow delivery

Security gates that add a week get routed around. Here is what works.

why your data team is firefighting article

Data & AI

Why your data team spends its week firefighting

Untested transformations and undefined ownership produce a permanent incident queue.

choosing boring technology article

Engineering

Choosing boring technology on purpose

Innovation tokens are finite. Spend them where they differentiate you.