Design a next word prediction system.

Focus on the architecture and algorithms that would enable effective next word prediction. Consider the types of data you would need to train your model, such as large text corpora, and discuss how you would handle preprocessing and feature extraction. Address the choice of machine learning techniques, such as n-grams, recurrent neural networks, or transformer models, and justify your selection based on performance and scalability. Additionally, think about user experience aspects, such as real-time prediction speed and context awareness, and how you would evaluate the system’s accuracy and effectiveness.