Understanding when and where events take place is important in fields such as intelligence analysis and law enforcement. Automated methods for mining events mentioned in a body of text can reveal mission-critical data patterns, hotspots, and trends more quickly. In this talk, we discuss solutions for bootstrapping a system which solves this problem with little data, explore the trade-offs it makes and lessons learned, and take a look at future research directions for scaling the model’s learning capacity and performance with larger datasets.