Data scientists and analytics engineers often model, integrate, and/or clean data within their data warehouse. Developers may need to use this prepared data to power applications. However, the services they build often need to do faster lookups than a data warehouse permits. To enable developers to move data from the warehouse into a low-latency, highly-reliable key-value store, Netflix developed Bulldozer. Bulldozer is a batch-based data platform that reads data from a data warehouse into DataFrames, converts each entry into a key-value pair defined in a protobuf file, and adds the key-value pairs into the key-value store. Because Bulldozer leverages a Key-Value Data Abstraction Layer, it can be used to write data to any key-value store (e.g. Cassandra, Memcached, etc.).