Tuning Systems: Experimentation for Engineers from A/B Testing to Bayesian Optimization

Does the program work as it should? Have changes made the system better or worse? Is it worth focusing on performance improvements instead of developing new functionality? Such questions can only be answered through experimentation. "Tuning Systems" will unveil the secrets of modern experimental practices developed and tested by world industries with the highest competition. The proposed practices will help enhance software development, machine learning systems, and quantitative trading.

The book provides a set of tools for optimizing programmable systems. You will journey from studying A/B testing to advanced experimental strategies that leverage the advantages of machine learning and probabilistic methods. The skills you gain from this practical guide will help minimize experimentation costs and quickly choose approaches and tools that will yield the best results for the business.