A Network Intrusion Detection System (NIDS) is an important tool that identifies potential threats to a network. Recently, different flow-based NIDS designs utilizing Machine Learning (ML) algorithms have been proposed as potential solutions to detect intrusions efficiently. However, conventional ML-based classifiers have not seen widespread adoption in the real-world due to their poor domain adaptation capability. In this research, our goal is to explore the possibility of improve the domain adaptation capability of NIDS. Our proposal employs Natural Language Processing (NLP) techniques and Bidirectional Encoder Representations from Transformers (BERT) framework. The proposed method achieved positive results when tested on data from different domains.