As Principal Engineer for Bitbucket, tackled complex challenges in performance, scaling, and data management for distributed git storage on AWS. Specialized in IO optimization, distributed caching, petabyte-scale data migration, and enhancing git robustness/performance at scale. Key achievements include significantly reducing NFS load via S3 offload, implementing dynamic IO throttling, optimizing storage costs through tiering, and architecting critical data lifecycle management features.

At NetApp, I built foundational expertise directly applicable to large-scale cloud storage challenges. As NFSv4 Performance Lead, I deeply optimized NFS IO performance (~40% gain) and state management. I tackled IO-intensive workloads like SAP Hana, developing the IOtrap library to boost throughput by ~50% via asyncio interception. My work architecting the distributed scale-out NAS (Infinite Volume) involved designing object storage components, high-performance distributed metadata search (BerkeleyDB, 5x faster queries), and performance benchmarking, providing critical experience in building and scaling complex storage systems.

At McAfee, I transitioned into a dedicated performance engineering leadership role, building the foundation for my subsequent work at NetApp. Responsibilities included building a performance team, establishing measurement practices, and hands-on optimization. I successfully identified and resolved key bottlenecks, such as reducing lock contention in multi-threaded code using custom memory allocators to achieve a ~25% scan rate increase. This experience, including practical application of tools like Intel VTune, was instrumental in preparing me to tackle advanced performance optimization problems in storage systems at NetApp.

Led the port of the Samba file server to VMS at HP, requiring implementation of a POSIX API emulation layer. This complex systems programming task provided foundational experience in file systems and cross-platform development.

At Bosch, I managed a team developing a data compiler pipeline and led its successful migration from Sun SPARC (Solaris) to SUSE Linux servers. This cross-platform transition was crucial for optimizing performance and reducing operational costs.