Practical thinking.
A closer look.
Notes on the decisions, disciplines and everyday work behind dependable technology.
9 posts
Artificial intelligenceHow to choose an AI project worth building
Start with a recurring business problem, a measurable outcome and a team prepared to own the result.
Artificial intelligenceAn AI evaluation your team can actually use
A repeatable set of realistic tasks helps turn a subjective demonstration into a reviewable engineering decision.
GovernanceBuilding an AI inventory that stays useful
A simple register with clear ownership can support better decisions than a comprehensive spreadsheet nobody maintains.
Cyber securityA practical first threat-modeling conversation
Bring the right people together, sketch the system and examine the assumptions that could expose your business.
Enterprise engineeringModernization without the big-bang migration
A staged approach gives teams room to learn while keeping the business processes they depend on in view.
Enterprise engineeringWrite the runbook before you need it
Operational documentation earns its place when another person can use it under imperfect conditions.
Strategy & deliveryWhat a good technology project brief looks like
A clear brief creates room for engineering judgment while making the business objective hard to lose.
Artificial intelligenceDependable AI starts with dependable data
Before introducing a model, examine the information it will depend on and the people responsible for keeping that information useful.
Strategy & deliveryThe project is not finished at deployment
A useful handover prepares the receiving team to make changes, investigate problems and own the system after the delivery team moves on.