During March 2025, blp@cs.stanford.edu developed Bloom filter persistence for the feldera/feldera repository, enabling Bloom filters to be written to and read from layer files. This work involved designing serialization and deserialization routines in Rust, leveraging data structures and file I/O to ensure filter state is preserved across system restarts, particularly when fault tolerance is enabled. By persisting Bloom filter data, the solution reduced reinitialization overhead and improved query consistency for FT workloads. The implementation demonstrated depth in system programming and performance optimization, addressing data integrity and startup reliability within the layer architecture without introducing new bugs during the period.