Written by the people who do the work
There is no content calendar here. These are working notes on problems we actually meet, published when we have something worth saying and something to lose by saying it badly.
8 pieces on engineering and applied AI.
Ordered as written. Each one is meant to survive you disagreeing with it.
Building practical AI workflows
When should businesses use AI agents?
Modernizing legacy applications
Designing SaaS for scale
API-first architecture
Human-in-the-loop AI
Cloud-native application architecture
Automating operational workflows
The questions underneath these articles.
Most pieces are a specific case of one of these larger arguments. Each links to the capability page where the work gets done.
Deciding where a model belongs
Plenty of operational decisions are deterministic. A rule that runs every time is cheaper, faster and easier to explain than an inference that occasionally disagrees with itself.
Applied AI and automationKeeping autonomy inside a boundary
An agent with a defined scope, a stopping condition and a named accountable person is a feature. An agent with none of those is an incident waiting for a quiet week.
Human in the loopModernising without a maintenance window
The system that cannot be switched off is the one that constrains every plan. Sequencing matters more than the target architecture diagram.
Legacy modernisationBoundaries between systems
Most integration pain comes from two systems each holding a different definition of the same entity. The API is the easy part of that conversation.
API and integrationWe publish judgement, not announcements
Product updates, new roles and company news belong on our own channels. What belongs here is engineering judgement: how we would approach a problem, what we would avoid, and where we think the industry is going wrong. No vendor comparisons and no engagement bait. If a piece is worth your time it should still be worth it after you have disagreed with half of it.
Turn a point of view into a scope.
The articles describe how we think. These pages describe what that looks like as an actual engagement.
Where the work happens
Each capability page covers fit, sequence and risks.
- AI and automation Bounded automation with evaluation.
- Software engineering Systems built around business rules.
- Cloud and DevOps Pipelines, infrastructure and observability.
The same argument, different constraints
Where compliance and access change the answer.
- Healthcare Consent and audit shape everything.
- Education Cohorts and academic years.
- Automobile Networks, bookings and parts.
Background
Who we are and what we build ourselves.
- About LivingTech One standard across products and client work.
- Folow Our platform, built to the same rules.
- Careers How we hire when no role is listed.
- Contact Bring a constraint, not a brief.
Reading is free. The conversation is more useful.
If one of these articles describes something you are dealing with, tell us the constraint. We will tell you what we would do about it.