Blog

Notes from the builders.

Things we have learned shipping AI, security, and cloud-native products. No buzzwords, just things we wish we had known earlier.

Postgres or a Vector Database? A Framework for Choosing

pgvector handles more than people assume, and a dedicated vector database handles less magic than the marketing suggests. Here is how we actually decide.

8 min·Jul 30, 2026Read

Cloud

What We Actually Use Kubernetes For (And What We Don’t)

Kubernetes gets reached for by default on projects that do not need it. Here is our actual decision process, informed by what worked at HostPepper.

9 min·Jul 10, 2026Read

Security

Why We Still Reach for Go for Security Tooling

Go is not the trendy choice anymore, but for the class of software that has to run everywhere with a small footprint and never surprise you, it is still the right one.

7 min·Jun 20, 2026Read

Picking Between Claude and GPT for a Feature: Our Actual Process

Not a vendor loyalty decision. Here is the evaluation process we actually run before committing an LLM feature to a specific model provider.

8 min·Jun 2, 2026Read

Shipping AI Features That Actually Compound

Why most AI features stall after launch, and the three product loops that keep ours improving every week.

7 min·May 12, 2026Read

Security

Zero-Trust Without Slowing Down Engineers

A practical guide to rolling out zero-trust networking on a real product team, without the politics and the painful developer experience.

9 min·Apr 28, 2026Read

Cloud

The Five Nines Myth: What 99.999% Uptime Really Costs

How HostPepper hit genuine reliability goals and why your product probably does not need five nines to deliver a great experience.

11 min·Mar 18, 2026Read

RAG Is Not a Feature. It Is a System.

A breakdown of the retrieval, ranking, and evaluation layers that separate production RAG from demo RAG.

14 min·Mar 5, 2026Read

Performance

How We Cut Page Weight by 46%

A case study in performance discipline: image budgets, edge caching, and the unglamorous wins that moved the number.

8 min·Feb 14, 2026Read

AI / UX

Designing Match Scores That Humans Trust

Showing confidence intervals, explanations, and counter-examples: the UX of AI ranking systems done right.

6 min·Jan 22, 2026Read

Subscribe for new posts

One note a month. No spam, ever.