A number of vital steps of the machine learning pipeline, including understanding the attributes of domain-specific data, defining prediction problems, creating a suitable training data set, and selecting a promising machine learning technique, tend to be done manually by a data scientist or a machine learning engineer on an ad-hoc basis. In this project, we are developing a Virtual Interactive Data Scientist (or VIDS), an intelligent agent, to remove the aforementioned bottlenecks and automate the entire machine learning pipeline as maximally as possible. VIDS is an intelligent agent that can formulate different prediction tasks and automatically recommend the best ones to its users.