Blog
Articles, notes and perspectives by Marco Zucco.
-
Building the team behind a high-scale platform
A platform is only as reliable as the team and the habits behind it. On building the people side of engineering at scale.
-
What I write about here
Marco Zucco on the focus of this blog — leadership, managing people, and putting AI to work in real business processes.
-
What pressure reveals about a team
Stress doesn't change a team — it shows you what was already there. On using hard moments to read an organisation honestly.
-
What reliability actually costs at scale
Every nine of uptime is paid for somewhere. In a high-traffic, regulated platform, pretending otherwise is how you end up paying more.
-
Delegation is a system, not a favour
Most delegation fails because it's treated as handing off a task. Real delegation transfers ownership — and that's a different discipline.
-
Architecture decisions that age well — and the ones that don't
Some technical choices get cheaper over time and some quietly compound into debt. Telling them apart early is most of the job.
-
What I look for before backing a founder
Investing early is mostly a bet on people. On the founder qualities I weigh most before putting capital and time behind a company.
-
Putting Claude to work: a practical playbook for AI in operations
Moving past chatbot demos. How we actually embed Claude into everyday business processes, and the rules that keep it useful.
-
Designing for the failure case, not the happy path
Most systems are built for the day everything works. The ones that survive are built for the day something doesn't.
-
Hiring for judgment, not just skills
Skills tell you what someone can do today. Judgment tells you what they'll do when the situation isn't in the playbook. On hiring for the latter.
-
The quiet cost of unclear ownership
When everyone is responsible, no one is. On the slow, invisible damage that ambiguous ownership does to an organisation.
-
AI won't fix a broken process — it will scale it
Automating a bad process just lets you do the wrong thing faster. Why AI rewards clarity and punishes mess.
-
One-on-ones that actually move things
Most one-on-ones are status updates in disguise. How to turn them into the highest-leverage half hour on your calendar.
-
What Claude's newest models changed in how we build
The latest Claude models aren't just better at answering — they're better at doing. A practical look at what agentic capability changes for a team.
-
Leadership under load: deciding with incomplete information
Waiting for certainty is itself a decision. On making good calls when you'll never have all the facts in time.
-
What actually breaks when a startup scales
Scaling rarely fails on the product. It fails on the things that worked informally and quietly stop working. On helping founders make the jump.
-
Building an AI-literate team
The advantage isn't access to AI — everyone has that. It's a team that knows how to use it well, and when not to. On building real AI literacy.
-
From prompts to processes: embedding AI in real workflows
A clever prompt is a party trick. Lasting value comes from building AI into a workflow that runs without you. On making the jump.
-
Culture is the behaviour you tolerate
Culture isn't your values on a wall — it's the worst behaviour you're willing to walk past. On what actually shapes how a team behaves.
-
Feedback that lands: saying the hard thing well
Most feedback fails not because it's wrong, but because of how and when it's delivered. On giving the hard message so it actually helps.