// Engineering notes
Deep dives on real-time infrastructure, Flutter architecture, AI agents, and lessons from production.
Why we moved our streaming platforms off managed RTC vendors — costs, SFU/egress architecture, token services, and the migration steps we run in production.
BLoC, repositories, dependency injection, folder structure — and the testing habits that survived production traffic across 15 Flutter apps.
The STT → LLM → TTS pipeline, latency budgets, and interruption handling patterns that make agents feel like teammates, not chatbots.
We share what we learn. If you have a topic you'd like us to cover, let us know.