EvanGruère

I like building software that still behaves when conditions stop being ideal. Most of the work I’ve enjoyed has had some kind of real constraint behind it: video pipelines where dropped frames are immediately visible, AI workflows that become hard to trust if you can’t see what they’re doing, data platforms where different roles need the same source of truth, and scraping systems that have to deal with rate limits instead of pretending the network is friendly. The domains are different, but the part that interests me is usually the same: what happens when the system is under pressure? I tend to think about problems from the architecture outward. Where does the data come from? Where does state live? What breaks first? What needs to be observable so that someone can operate the system with confidence? Once those questions are clear, the choice of tools becomes secondary. I’m comfortable moving between TypeScript, Rust, Python, and mobile work when that helps the system make more sense. A recurring theme in my work is control. Not control for its own sake, but the kind that makes software easier to understand, debug, and evolve. I’m drawn to systems where the important behavior is explicit rather than hidden behind black boxes. That has shown up in different forms: building workflow engines for AI interactions, working on self-hosted mapping infrastructure, and designing real-time streaming systems. In each case, I care less about using the newest tool and more about knowing how the pieces fit together when the product is running for real. Lately, I’ve been exploring this through projects around sovereign data and self-hosted infrastructure, especially in geographic systems and travel. I’m interested in what becomes possible when products are not shaped entirely by external platforms. I like problems where the constraints are real, the trade-offs matter, and the system has to keep working over time — not just pass tests.