Kingston, Kingston, Jamaica
Student researching and implementing lock-free concurrent algorithms at the intersection of high-performance and modular database storage engines. Currently building a latch-free, log-structured storage engine in Rust based on the LLAMA architecture (Levandoski et al., VLDB 2013).
Life is too short. Why not make a Database? BloomDB is a ground-up database engine built layer by layer, starting from a latch-free storage core and designed to bloom upward, layer by layer, into a full database system. The name reflects the philosophy of the project. Each layer is a self-contained abstraction that composes cleanly onto the one below it, following the Deuteronomy decomposition principle. Currently implementing LLAMA, a latch-free, log-structured storage engine designed for modern multi-core hardware (Levandoski et al., VLDB 2013). Highlights so far: • Fully cooperative latch-free write path achieving 4–4.5 GiB/s write throughput in synthetic stress tests • Log-structured storage layer with fully asynchronous I/O •Architected for composability. Storage and caching concerns fully isolated from transactional logic, following the principle of decomposing the database kernel.
• Led development of an automated internal tool for loan record retrieval using Power Apps, SharePoint, and SQL Server, transforming a manual paper-based process into a cloud-based solution, targeting a reduction in authorization time from ~2 days to <24 hours. • Collaborated with supervisors and end-users to gather requirements and iteratively refine system design based on employee feedback. • Provided hardware/software troubleshooting for 50+ employees, minimizing downtime and sustaining cross-department productivity