I work on problems where technical decisions have economic consequences.

My background is engineering and industrial systems, spanning from design, structural and fluid simulation to large-scale manufacturing. That work taught me to think in constraints: cost, feasibility, scale and trade-offs rather than ideal solutions.

Today, I focus on optimization, analytics and machine-learning systems, with a particular interest in cost- and energy-aware architectures. I care less about individual techniques than about how systems behave as a whole, where complexity is necessary, where it isn’t, and how decisions compound over time.

I'm drawn to problems at the intersection of technology integration, quantitative analysis, and strategy, especially in infrastructure-heavy or capital-intensive contexts. My goal is to make trade-offs explicit and decisions easier to reason about.

I’m reducing RAG energy consumption at inference time, focusing on how dynamic architectures and activation choices affect energy consumption and latency. My work involves system-level energy measurement and static configurations replacement with ML-logic conditional activation that adapts computation to the query.