The use of underground space for engineering systems has been increasing worldwide. Underground geoengineering is characterized by complex, uncertain geology and geomechanics that present challenges and require new tecniques to be dealt with. These challenges include large overburden which cause high stresses and temperatures, that require complicated engineering design. Additional environmental challenges exist in cases related to petroleum engineering, nuclear waste disposal, storage of products and energy, storage of CO2, geothermal energy and others. In large projects, typically a large amount of geotechnical data is generated. This data can hold valuable information that can be used to improve decision making and optimize design and construction processes. It is therefore necessary to define standard ways of collecting, organizing and representing the obtained data. There are automated Artificial Intelligence (AI) tools and pattern recognition techniques that enable one to analyze this vast data - Data Mining (DM) techniques. After discussing the general challenges associated with deep underground engineering and the application of AI and DM techniques in underground construction, this paper presents case studies where innovative DM-based techniques were developed and applied by the authors. In particular, the paper demonstrates the application of DM to the design and construction of a large and deep underground hydroelectric scheme, an underground laboratory, and undergound mining specifically based on rockburst risk assessments.