Sep 2024 – Feb 2026 · 1 yr 5 mos · Menlo Park, California, United States · On-site

• I support the AI Training Systems team at Meta. We build the scheduling, orchestration, and data processing systems that power some of the largest-scale AI Training in the world. We're responsible for building reliable infrastructure, tools, and services that are used every day by ML engineers and researchers across the company.

Feb 2021 – Jun 2023 · 2 yrs 4 mos

• Led an organization of 180 engineers and managers responsible for enabling Meta's infrastructure to support the growth and diversity of products such as Facebook, Instagram, WhatsApp, and Threads.
• I supported multiple teams and stakeholders building Meta's cluster management system, Tupperware/Twine. Tupperware is Meta's version of Kubernetes/Mesos responsible for container allocation and lifecycle management for Meta’s tens of thousands of backend services.
• The team developed and ran the software responsible for the infrastructure lifecycle operations (provisioning, configuration, maintenance, etc.) of Meta's expansive fleet, which consisted of over ten million hosts across 20+ datacenters.
• The team also owned the operating system, packaging, deployment, and host configuration systems for all Meta services.
• Pioneered a program to hire candidates from nontraditional backgrounds (e.g., self-taught individuals, coding boot camp graduates, career re-entrants) into a year-long immersion program.

Jan 2015 – Feb 2021 · 6 yrs 1 mo

• Established and led new PE/SRE teams within the Online Data organization, which housed Meta’s backend data stores and caches, each handling billions of requests per second.
• Led the Production Engineering/SRE teams for TAO (Graph write-through cache), Memcache, ZippyDB (Distributed key/value store), and graph indexing systems.
• I drove a Reliability initiative that significantly improved the reliability of Meta's online data systems. I led the weekly incident review, participated in an incident manager on-call rotation, and facilitated cross-functional collaboration to improve system reliability.
• Orchestrated backend service migration from bare metal to containerized environments.

Sep 2012 – Jan 2015 · 2 yrs 4 mos

• Designed and implemented a datacenter maintenance orchestration system, automating the process of shifting backend service capacity between datacenters.
• Led capacity and disaster recovery planning for Facebook's distributed data caching layer (TAO and Memcache), including the creation and implementation of operational tools.
• Developed a reusable monitoring framework to standardize alerting and remediation procedures across multiple backend data and cache services.
• Engineered an on-call paging system that efficiently looped through on-call personnel, automating call and escalation procedures. This system was deployed company-wide and enabled the phase-out of a 24x7 operations team.

Aug 2011 – Sep 2012 · 1 yr 1 mo · Brandon, MS

• Supports data center technical services comprised of varying local and wide area networks, email delivery systems, data warehouses, directory services, web infrastructure, and over 200 servers. Develops various enterprise-class systems for enhanced service delivery. Supports high-complexity network backbones in nearly 30 schools across the Jackson, MS Metro Area.

Jan 2009 – Dec 2009 · 11 mos · Tuscaloosa, Alabama Area

• Developed mission-critical, internal applications for use by University health officials. Designed websites for University Health Services and the Longest Student Health Center.