I am glad that you liked the article and are interested in the upcoming course :)
In general yes, of course — you can participate and implement the projects of the course in Keras. Although Keras may cause some problems when you are going (if you want) to deploy the finished models to production in the last step of the projects. But we can deal with that.
If you prefer Keras I would nevertheless suggest you not to use “vanilla” Keras but rather using TensorFlow Keras API, which incorporates all functionalities of Keras, while being a part of TensorFlow. By using this API the code remains simple but as robust as regular TensorFlow code.
TensorFlow as well as Keras API will be topics of the course also, btw ;)