In IDA, we try to make computers discover knowledge in data. To this end we develop data mining and machine learning algorithms helping us detect regularities (frequent patterns, strong associations), construct predictive models, and ultimately identify the phenomenon that generated the observed data. Our expertise includes the strong paradigm of relational data mining, overcoming the deficiency of traditional data mining tools which constrain their analysis to a mere single table of a multi-relational database. Our prominent application areas are in bioinformatics (mainly genomics) and product engineering. Our second effort stream is concerned with intelligent optimization. Here we devise unorthodox techniques, such as novel evolutionary computation or randomized search algorithms, providing reasonable solutions where traditional methods fail. While data analysis problems generate numerous optimization tasks, optimization can conversely benefit from data analysis techniques (e.g. by learning from old problems' solutions). Our long-term vision is thus a hybrid intelligent system mutually bootstrapping both paradigms -- an idea we term adaptive optimization.