As service/product provider you want to make fraudsters sweat for every cent, without affecting legitimate users. Fraud occurs every minute of every day, it is a matter of volume to get you to do something about it. Traditional fraud prevention involves business rules, however today the main approach is to use AI, Machine Learning algorithms for this. In Gett, we found that we must have both sides, as algorithms need a large amount of data to be effective. I will focus on Rule Engine implementation, describing how you can decouple rule writing from rule execution and show a rule engine written in JavaScript, executed inside GoLang Service.