INSIGHTS

What we’re learning, in the open.

Field notes from building and shipping AI-native software in regulated enterprises. Straight from the actual work.

Apr 2, 2026Agentic Engineering20 min read

Why agentic AI projects fail in production

Most agentic projects fail in production not because the models are bad, but because teams try to prompt-engineer their way out of writing actual control flow. The real failure modes, and the architecture that works instead.

Vijay Swami

Mar 26, 2026Healthcare · Building6 min read

Lessons learned from shipping a regulated healthcare app with Claude Code

What actually got built shipping enterprise healthcare software with AI coding tools, and the hard-won lessons from production.

Vijay Swami