Articles on AI-native development, LLM integrations and AI agents, Flutter, web and backend engineering, plus lessons from the products we build and run.
We built Claude-based agents for three departments of a construction company in Iași — tenders, civil works and hydro — and trained the team to use them, improve them and build new ones. What the agents do, how the training worked, and the rules that keep people in charge.
Read more →On a native iOS + Android project, a developer builds each feature on iOS with AI assistance; once it is merged, an agent ports it 1:1 to native Android, a second agent reviews the port, both apps are tested side by side, and a senior engineer reviews and tests last. How the pipeline works and what each layer catches.
Read more →Calling a vision model is the easy part. ParsePrime turns PDFs and photos into CSV tables for paying customers, and most of the work went into what surrounds the model: output contracts, untrusted documents, cost ceilings, safe files and tests that run without the model.
Read more →After 5 years of Flutter in production and one in-house Flutter web product (Table Online), an honest assessment: when Flutter web is the right choice, when it isn't, and what edge cases to avoid.
Read more →Lessons from building Table Online: a real-time multiplayer game, admin dashboard and backend in Dart, plus a small Go realtime service. Why ‘Flutter = mobile’ is a limitation we outgrew with a small team.
Read more →Engineering lessons from building a real-time game in Flutter, with the same code running in the browser and natively on mobile. State management, server sync, 60fps performance.
Read more →
