I have spent years listening to machines.
Metrics, traces, logs, alerts —
small signals rising from the deep water of distributed systems,
asking to be understood.
Along the way, I began building systems for agents —
harnesses, AI-native platforms, and the infrastructure
that lets them navigate codebases, use tools, and follow workflows.
Today, I work on LLM evaluation —
designing coding and long-horizon tasks,
exploring what models can do, where they fail,
and how the harness around them shapes their behavior.
Somewhere between infrastructure and intelligence,
I am still doing the same thing:
turning chaos into language, and language back into systems.