These pages are about fast data science, or using languages and techniques outside of the the conventional R/Python stack to improve the performance of computer models that do machine learning, optimization, simulation, inference, or other complex tasks.

This journey started with a number of models that I translated from Python into Julia and Go, and achieve huge speed-ups that not only reduced the time to find answers, but also opened up new possibilities for using these models, such as putting them behind web-based interactive front-ends, allowing people to use the models in new and better ways that would not have been possible with standard approaches.

More recently, I have been applying Rust and Zig to data science and AI problems, and am also exploring the use of GPUs for computationally-intensive models.

I am Andreas Kaempf, Accenture’s Managing Director and Head of the Global Centre of Excellence for Agentic AI, for Central and Eastern Europe, based in Munich. My role is to advise clients and teams on generative and agentic AI. I am also the EMEA-wide lead for AI applications architecture, and advise on other types of AI such as machine learning, reinforcement learning, optimization, and simulation.

This is a personal web site, and the views expressed here are my own.