Skip to content
CloudMindSolutionsCloudMind Solutions Inc.
Insights

Notes from the work, not from the marketing calendar.

We write when we learn something worth passing on — usually after an engagement surfaces a problem we had not seen framed properly anywhere else. No posting schedule, no SEO filler.

posts
4
cadence
When there's something
written by
The engineers

Showing 4 posts.

Applied AI7 min

Your RAG system's biggest risk isn't hallucination. It's permissions.

Everyone tests retrieval systems for made-up answers. Almost nobody tests them for correctly-sourced answers shown to the wrong person — which is the failure that ends up in a breach notification.

[PLACEHOLDER] Founder Name

Cloud6 min

What a dependency map costs — and what skipping one costs more

Two weeks of discovery feels like two weeks of nothing happening. It is the cheapest insurance available on a migration, and the reason most overruns are decided before any workload moves.

[PLACEHOLDER] Founder Name

Security6 min

Ninety-one findings is not a security report

A list sorted by CVSS score is not a priority order, because severity is a property of a vulnerability and risk is a property of your network. Here is what to ask for instead.

[PLACEHOLDER] Founder Name

Software7 min

Characterization tests: how to change code nobody understands

You do not need to know why a behaviour exists to protect it while you work around it. This is the technique that makes legacy modernization survivable — and the one most teams skip.

[PLACEHOLDER] Engineer Name

Recognise one of these problems?

Most of what we write about started as a client engagement. If a post describes your situation more accurately than you'd like, that's usually a good reason to talk.