The application of autonomous driving has led to a surge in computation-intensive and delay-sensitive vehicular applications. These crucial applications typically generate a large amount of data that must be processed with low latency, posing a significant challenge for local on-board processing due to the usually limited on-board computing capabilities. To alleviate the computation burden, mobile edge computing (MEC) has emerged as a promising approach to provide additional computing power with low processing delay. Simultaneously, the high demand for data sharing among vehicles can be facilitated through vehicle-to-everything (V2X) communications. This talk explores the key resource allocation problems of MEC by deploying it in the vehicular scenario. Computation-intensive vehicular applications are modeled in a latency minimization problem, enabling the adoption of proper allocation schemes to improve latency performance. Finally, it outlines potential future research directions in the field of AI applications for solving the complex resource allocation problems in MEC systems.