About
I'm Cloves Almeida, a Principal AI Engineer working on production AI systems for large enterprises. This site is a place to write technical notes about the parts of applied AI that don't fit into a model card: decision layers, constraints, evaluation harnesses, optimization, monitoring, and the messy interfaces between research and production.
My background is twenty-plus years across software and AI engineering, with recent work spanning pricing, scheduling, allocation, and decision optimization for Fortune 500 clients. I have a NeurIPS-published applied method developed during a client engagement, benchmarked against state-of-the-art baselines, and deployed in production. Earlier in my career I competed on Topcoder, and I still think most production AI judgment starts with good algorithmic instincts.
The notes here are deliberately sober. They are not thought leadership. They are closer to engineering memos: short pieces on recurring failure modes, system decompositions, and tradeoffs I've found useful when building AI systems that have to run inside real operational constraints.