In previous chapters, we explored the classical foundations of causal inference [1]: potential outcomes, interventions, and methods to estimate causal effects. These methods are powerful but are often designed for relatively simple, low-dimensional datasets where manual adjustment is feasible. However, modern data, such as images, text, and sequential data, is increasingly high-dimensional and complex. In these settings, traditional causal methods face significant challenges. Meanwhile, deep learning has emerged as a transformative technology capable of learning rich, non-linear representations from raw data. This chapter introduces how deep learning can be integrated with causal inference [1], the challenges involved, and the specialized models developed to bridge these two fields.