Kenyi Takagui-Perez

I am an applied AI engineer at Yavendio, where I work on agent architectures and workflows for sales.

Before that, I was a machine learning engineer at Yape, building recommendation systems; at Spatialise, a Dutch startup using graph neural networks to predict soil organic carbon at high resolution; and at INRAS-PUCP, where I worked on inverse algorithms for electron density profiles and neural networks for electron-trace retrieval.

Earlier still, I studied physics in Peru at PUCP and in Argentina at the Balseiro Institute, with stops along the way in condensed matter, particle physics, quantum information, and electronics. My master's thesis studied the effect of Coulomb repulsion on Majorana zero modes in quantum-dot–superconducting-nanowire hybrid systems, supported by a CNEA Fellowship.

These days, I am mostly tinkering with agent architectures and LLMs. I also keep up personal projects in AI infrastructure and ML for materials science and biology.

I am originally from the highlands of Peru and spent some good years in Argentine Patagonia. In my spare time I enjoy cycling; playing guitar, cello, and piano; and exploring other forms of art. Lately I've grown curious about the investment world too.

Publications

Find all published research here.