Species83 is my company focused on the development of homelab hardware platforms for distributed computing and storage.

I am leading the research and development efforts for Clyso GmbH in North America. We are primarily offering boutique services and open source solutions for HPC, distributed storage, and cloud, with a specific focus on Ceph.I am also staying on as the performance lead for the upstream Ceph project and still attend the weekly leadership meetings. I will also continue to run the weekly community performance meetings as before. Please feel free to join us: https://ceph.io/en/community/meetups/Stay tuned for more!

I am one of the core engineers for an open source distributed storage system called Ceph. My work primarily focuses on performance of our underlying storage daemons, though I also consult on hardware architecture and benchmarking. Most of my work is in C++, though occasionally I write specific projects in other languages depending on the use case. Examples of my recent projects include:- Memory usage autotuning and automatic cache balancing via age binning.- LRU Cache trimming and locking code refactor for ~30% performance gains. - A Wallclock profiling tool based on libunwind and elfutils for performance and lock contention analysis. Fast (~145K samples/second) remote unwinding without parasitic injection.- The Ceph Benchmarking Toolkit (python based).- RocksDB and other Key/Value store optimization and analysis.- A go based concurrent S3 object put/get/list/delete benchmark.- I run the weekly ceph community performance meeting.

At Inktank my roll was to analyze and understand everything that affected the performance of Ceph (An open source distributed storage system developed by Inktank). I benchmarked code, debugged kernel issues, wrote whitepapers, built tools, and designed hardware platforms.

My roll at MSI evolved from software development, to HPC system administration, to project leadership. Some of the projects that I worked on included:- Building a prototype openstack cloud for HPC research.- Writing tools to analyze cluster utilization, performance, and job scheduler efficiency.- Co-Leading development of a proteomics analytics platform.

I joined a research and development team at General Dynamics AIS to create a network-centric command and control platform written in Java and C++. I built a complete set of network APIs necessary for this platform to interact with existing army tools such as FBCB2.

I maintained the IT infrastructure for the Center for Cognitive Sciences and several research laboratories in the Department of Psychology. This included various Linux and Unix based web, mail, and file servers along with Mac and Windows desktop clients.