Brandon S.

- Team Lead of 4 engineers - Built a Gitlab CI/Argo CD model supporting simultaneous deployment to multiple Kubernetes clusters across different cloud providers (Linode LKE, Amazon EKS). - Ensured 99.99% infrastructure uptime across two prod environments. - Architected automated promotion of code across environments. - Utilized Infrastructure-as-Code to create persistent and ephemeral Kubernetes clusters. - Standardized Kubernetes manifests by writing a common template library in Helm. - Modeled clusters through Helmfile/Go templating, Cuelang + Timoni, and more. - Created a full monitoring and alerting stack from the ground up using Datadog (later, as code). - Implemented ChatOps tools, such as a bot giving readonly access to k8s clusters through Slack. - Authored an Ansible playbook to manage a fleet of Gitlab runners. - Created a follow-the-sun on-call rotation which allowed engineers to only be on call during their daytime, using OpsGenie. - Configure kafka, redis, and cloud-native posgres clusters for our applications - Create RBAC to internal resources via Teleport/Tailscale VPN - Configure SSO across all services via Okta

- Reduced CI pipeline run time by 85% through parallelization and custom Jenkins shared library code. - Automated the entire CD process, condensing a legacy 6-pipeline solution (requiring input at every stage) down to a single button click. - Transitioned CI from a UI implementation to a declarative (code) approach. - Added vulnerability scanning and other security gates across all teams’ pipelines. - Converted Kubernetes manifests into fully templated helm charts. - Automated the schema migration process. - Responsible for 2 clusters; a staging cluster and a production cluster. - Ran support on-call for our hosted SaaS and On-Prem customers.

- Architected caching and state layers for our backend. - Created a modular persistent storage solution, allowing the backend to connect to many database technologies. - Wrote service management components and REST controllers in Java 8 / Spring Boot / Gradle. - Developed a custom metrics collection and analysis tool suite, used to create timing guarantees for our backend. - As scrum master, chieved 25% velocity boost for each engineer on the team - Reduced Bamboo CI build times by up to 4x. - Containerized build environments with Docker. - Implemented static code analysis using SonarQube. - Led an initiative to migrate from an on-prem dev environment to a hybrid cloud solution.

I was selected as a participant for the NSF IRES grant. This grant sent me to Chalmers University of Technology in Gothenburg, Sweden, where I began my graduate research. My research was focused on identifying bottlenecks within the Linux storage stack for HDDs, SSDs, and RAM for real-time and embedded systems. I have used both a custom tracing framework for in-depth function tracing at specific layers, and pre-existing tools (trace-cmd/ftrace/eBPF) for profiling. The framework I built modified to the filesystem and block layers of the storage stack, a workload generator, a /proc module for trace toggling (modeled after ftrace), and my own analysis toolkit. My thesis identified several bottlenecks in the storage stack and details a way to build a kernel tracing framework from scratch when a custom profiling tool is needed.

I was offered to do research for a semester at Chalmers on an NSF IRES grant, where I began formalizing my master's thesis under Dr. An-I Andy Wang with help from Chalmers faculty member Dr. Sally A. McKee. During my stay, I learned about the Android and Linux storage stacks. Notably, I used ftrace to profile the Linux storage stack, identifying some high-latency functions of interest within I/O data paths.

Undergraduate research under Dr. Andy Wang and PhD student Mark Stanovich as part of the Operating Systems research team at Florida State University. I wrote programs to bombard machines with network packets and analyze target machines, which used various real-time scheduling algorithms to see how workload processing was effected.

I serviced help desk tickets related to user accounts/server problems, automated certain tasks (such as password generation/recovery), and migrated the mail server to dovecot.

Bachelor of Science in Computer Science, with minors in mathematics and physics.