The systems problem underneath the craft
Brewing is applied biology and chemistry with a tight feedback loop. You set process variables, run the fermentation, measure the outcome, and adjust. The recipes that work are the ones where you understand why they work, not just that they do.
That framing is where my interest sits. Not the romanticism of craft brewing, but the repeatability problem: how do you get consistent results from a process with dozens of variables and a multi-week cycle time?
What I brew
The range covers traditional beer styles (IPAs, stouts, Belgian ales), dry-hopped ciders, and experimental ferments including mead, kombucha, and lacto-fermented hot sauces. Each is a slightly different process control problem.
The data angle
Recipe iteration without data is just guessing. I track fermentation temperature curves, pH at multiple stages, final gravity, and sensory notes across batches. The Brew Codex project started here: I wanted a structured dataset I could actually query across batches, not a collection of notes.
Water chemistry, yeast selection, and mash temperature decisions all affect outcome in ways that compound. Tracking them systematically is the only way to know which variable moved the needle.
What it connects to
The same instincts that make fermentation interesting make architecture interesting: understand the system before tuning it, build in measurement from the start, and treat the first batch as a learning instrument, not a finished product.
What is next
Building a fermentation temperature control system with sensor integration. Extending the recipe database to support cross-batch analysis. Eventually exploring whether ML has anything useful to contribute to hop substitution decisions.