CV
For a downloadable version, get in touch.
T. Michael Cornelia
Athens, GA · LinkedIn · Contact
Data center operations Director responsible for ~24% of Meta’s GPU fleet and power capacity. 13 years scaling Meta’s fleet from megawatts to gigawatts; deep operational ownership of GPU training and inference platforms — Zion, Grand Teton, SMC, MTIA, and NVIDIA GB200/GB300 GenAI clusters. Founded and led Meta’s data center AI Operations program — drove a 49% reduction in unplanned downtime and interruptions on training infrastructure and 70% on inference infrastructure in year one. Engineer → Director — engineer’s instincts paired with director-level strategic range.
Selected Impact
- Regional growth: Scaled South region 754 MW → 1,070 MW (+42%) in 18 months
- Reliability: Held South region unavailability below targets on training and inference platforms through a 6x scope increase (2025)
- Capacity throughput: Landed and provisioned 2,307 racks in 2024 at 2.2-day average (vs. 3.7-day fleet); 98.8% SLO (vs. 93.8% fleet)
- AI Operations program: Led a 49% reduction in training downtime/interruptions, 70% reduction in inference downtime, 180% improvement in diagnostic accuracy (2022)
- People: Top-quartile manager engagement scores (2025); promoted 3 individual contributors to Site Manager (2024-25)
Experience
Director, Global Operations — South Region
Direct all site operations for Meta’s South region — ~24% of the global GPU fleet and total megawatts. Scaled the region 754 MW → 1,070 MW in 18 months (+42%) including a 500+ MW generative AI training cluster; on path to ~2,500 MW by 2027. Currently supporting the turn-up of Hyperion in Louisiana — Meta’s next-generation campus, the size of Manhattan and 5 GW at full build-out.
AI & GPU Infrastructure Operations
- Founded and lead Meta’s data center AI Operations program (2022) — drove 49% reduction in unplanned downtime and interruptions on training infrastructure and 70% reduction on inference infrastructure in year one; improved diagnostic accuracy 180%
- Operate GPU training and inference platforms across the South region; established the AI operations knowledge base, dashboards, and weekly leadership cadence adopted as the source of truth across the data center org
- Co-created the new hardware introduction process used for Grand Teton, MTIA (Meta Training and Inference Accelerator), NVIDIA GB200, and GB300 deployments
- Authored the unplanned-downtime alerting strategy and SEV0 incident response plan, institutionalizing incident response across the site operations org
- Launched cloud operations pilot across multiple public cloud providers (OCI, GCP), defining the support model for heterogeneous cloud infrastructure
New Region Turn-Up & Commissioning
- Delivered Meta’s first all-Turin server region on schedule — proving an accelerated capacity program targeting 80% reduction in fulfillment time
- Positioned 2,063 racks at rates up to 750 racks/week, single-day record of 200 racks
- Manage continuous turn-up pipeline across liquid-cooled facilities, rapid deployment structures, and leased data centers; 12+ concurrent new builds delivering 150–330 MW per quarter
Organizational Strategy & Workforce Planning
- Authored and globally deployed the Leadership Deployment Model — 9-month effort with HR and Legal enabling site operations to scale 3x (to ~10GW) without adding regional leadership headcount
- Led convergence of facilities and site operations risk frameworks into a unified Data Center Operations metrics dashboard — first shared operational data layer between the two orgs
- Co-authored ring-aligned restructuring proposal — shifted operations from per-site to infrastructure-ring-based model; 34% management reduction while scaling to gigawatt-class campuses
Capacity Planning & Risk Management
- Created region-level capacity delivery risk framework combining construction risk with operational signals into a unified executive dashboard
- Standardized capacity engineering processes, launched root cause corrective action (RCCA) framework, and established quarterly quality assessments
- Manage region against rack turn-up SLO (P90 < 5 days), redeployment SLO (P95 < 4 days), and decommission targets with per-site tracking
Government Relations & Community
- Represent Meta in state-level legislative advocacy for data center tax incentive preservation
- Graduate of Leadership North Carolina; maintain statewide network of leaders across public and private sectors
- Manage relationships with government officials, economic development authorities, and community organizations across eight states and growing
Director, Site Operations — Stanton Springs, GA
Directed site operations for Meta’s Stanton Springs (Newton County, GA) data center campus — one of the fastest-growing campuses in the fleet.
- Sponsored enterprise-wide Quality Metrics Workshop — multi-tier composite scoring framework retired legacy metrics; introduced LLM-based quality assessment of repair tickets
- Built operational excellence framework (1:1 templates, meeting cadences, analytics SOPs) adopted across all regions globally
- Launched diagnostic accuracy initiative — reduced repeat repairs, slowed headcount growth, and quantified cost impact to justify tooling investments
- Led insourcing business case analyzing contingent vs. FTE economics across 500+ positions, including staffing models for gigawatt-scale sites through 2030
- Built government and community relationships across Newton County and the state of Georgia
Data Center Operations Manager — Stanton Springs, GA
Managed infrastructure operations teams at Meta’s Stanton Springs campus during a period of rapid expansion and new building commissioning.
- Owned all capacity-related projects across the campus as it went from dirt to provisioning, including new building commissioning and expansion phases
- Directed the turn-up of the largest known NVIDIA A100 GPU neural network at the time; followed by the largest known H100 GPU neural network
- Managed cross-functional coordination across construction, production operations, and headquarters teams for turn-up, turn-down, and retrofit execution
- Developed operational processes and standards subsequently adopted at other sites
- Mentored individual contributors across multiple technical teams
Data Center Operations Manager — Forest City, NC
Led infrastructure teams at Meta’s North Carolina campus during rapid fleet expansion.
- Owned all capacity-related projects across the campus, including commissioning of the site’s third data center building
- Managed cross-functional coordination across construction, production operations, and headquarters teams for turn-up, turn-down, and retrofit execution
- Developed capacity processes adopted fleet-wide; mentored individual contributors across multiple teams
- Built local government and community engagement relationships at city and county levels
Earlier Career
Systems Engineer & Architect — SUM/IT Systems · 2004 – 2013
Designed and deployed systems solutions (Unix, Linux, VMware, Solaris, Windows Server) for SMB customers. Co-created the company’s cloud offering and managed full customer lifecycle from scoping through long-term support.
VP, Operations & Information Systems — The School Box, Inc. · 2000 – 2013
Directed technology operations for a multi-state retailer (~400 employees). Managed all infrastructure, partnered with CEO/CFO on strategy and budgets, and built the organization’s primary technical strategy and knowledge management platform.
Core Competencies
Data Center Operations · Hyperscale Infrastructure · GPU Fleet Operations · AI/ML Training Clusters · New Site Commissioning · Liquid Cooling · Capacity Planning & Delivery · SLO Management · Agentic AI for Operations · Multi-Cloud (OCI, AWS, GCP) · Colocation Management · Workforce Strategy · Organizational Design · Quality Systems · Operational Excellence · Legislative Advocacy · Government Affairs · Community Development
Education & Certifications
Leadership North Carolina — Statewide leadership program (graduate)
University of Georgia — Business Management coursework
Red Hat Certified Engineer (RHCE)
Red Hat Certified System Administrator (RHCSA)
Six Sigma Green Belt