Insights

Engineering notes from real builds

Practical writing on AI systems, production architecture, data infrastructure, and the tradeoffs behind software that needs to keep working.

All insights

Lessons from systems already running

Each of these came out of a system that was already in production. Where we changed our minds, we say so.

Hiring a studio

What to ask before you pay a web agency

Five questions that separate a studio that will finish your project from one that will take a deposit and go quiet. Ask them of us too.

Read

Performance

Why your website is slow, and what it costs you

Slow sites lose customers before the page finishes loading. The causes are usually the same four things, and none of them require a redesign to fix.

Read

Ownership

What you actually own when your website is finished

Domain, hosting, code, analytics, and content. If any of them are in your agency’s name rather than yours, you have a problem you cannot see yet.

Read

AI Engineering

Why most RAG systems fail in production

Retrieval augmented generation looks simple on paper. In practice, naive implementations surface irrelevant chunks, hallucinate citations, and degrade under real query distributions.

Read

System Architecture

Designing reliable multi LLM systems

Single model pipelines break under edge cases, rate limits, and model drift. A robust multi LLM architecture introduces fallback chains, output validation, and cost routing.

Read

Healthcare AI

Building healthcare AI systems responsibly

Healthcare AI carries higher stakes than most domains. HIPAA compliance, ICD-10 terminology accuracy, and clinical validation introduce constraints that pure ML thinking ignores.

Read