David Butler Email & Phone Number

• Solved critical resource exhaustion through establishing centralized logging with Grafana/Loki which provided comprehensive view of system usage and health• Developed Docker-based deployment option for flagship AI/ML product, including building related infrastructure and CI/CD pipeline from scratch. This facilitated migration from Heroku to AWS EC2 and also enabled dark-site customer deployments• Increased addressable market by adding Snowflake as an additional back-end and restructuring code to make additional integrations possible

San Francisco Bay Area

• Collaborated locally and remotely with different teams and stakeholders: coordinated with product managers to gather requirements and write technical designs and RFCs to meet emergent customer demands, acted jointly with QA and customer support teams to ensure that new features were properly documented and tested, and provided support for sales engineers. Tracked on-going issues on JIRA, and often worked asynchronously with teams in different time-zones • Ensured seamless and reliable customer experience by investigating root causes, troubleshooting, and resolving bugs across the stack, including: Linux kernel, high performance core proprietary system services, and cloud infrastructure in a multi-threaded, multi-process, multi-host distributed environment. Mined historical logs for statistics and then presented graphical reports for area leads and management to provide evidence based communication• Migrated existing AWS cloud product to Azure, allowing expansion into new markets. Integrated upstream Linux kernel patches in C and developed new patches to optimize product for Hyper-V hypervisor, for IO/CPU scheduling, and for Accelerated Networking support. Performed and analyzed benchmarks using customer-like workloads to ensure consistent system performance• Led the design and implementation of an internal diagnostic service for cloud deployment monitoring, which provided a fresh, system-wide view of outages, resource consumption, and activity. This allowed previously disparate information to be easily time-correlated and localized, greatly reducing time required to root cause complex issues in a distributed environment• Enabled new distributed use cases through additional IP based access control features. Owned feature implementation from end-to-end, including: in-kernel iSCSI server, core data-plane services, control-plane, CLI, and Rest API

• Deployed, administrated, and scaled critical infrastructure as the company grew, including servers running on bare-metal and on EC2. Designed security policies and firewall rules. Set up backups and replication. Sand-boxed application services and resources of differing concerns through containerization• Built minimal Linux based digital signage appliance OS, using Buildroot, C, C++, and OpenGL. Efficiently utilizes network and compute resources so that it can run on low cost hardware (Raspberry Pi, other misc SBCs and various mini PCs) with limited and/or inconsistent internet access. Includes automatic software updates and configuration management• Created 100x improvement in used disk space and query time compared to previous PostgreSQL database based ad play accounting solution. Achieved through architecting and enacting a bespoke service that reduces overhead by compressing and summarizing time blocks• Engineered a new digital signage management system. Implemented role based identity and access control system to secure media and device control in a multi-user environment. Implemented web based front-end using HTML and JavaScript, with a back-end using Python, Django, and PostgreSQL• Led monitoring of on-going, incremental, non-disruptive software updates to distributed fleet of remote signage players