 Created supply planning process in google sheets that has been used for the last 2.5 years to automatically handle resource demand requests (cores, terabytes, servers, etc), convert to specific hardware types and quantities based on a hardware roadmap, add predefined buffers and then output a final server rack forecast. This process also automatically determines the server fleet refresh demand needs based on what is allocated on the datacenter floor and when refreshes are planned. After server rack orders are defined, the process auto-designates them to a specific datacenter to ensure no stranded power. The power usage of the forecasted server racks is then used to determine a 3+ year rolling build/decom/retrofit plan for our site selection and sourcing teams to contract out. Infra level strategies developed and approved:- How/when we forecast, plan and lock our rack orders with respect to demand changes, supply chain lead times and datacenter space and power availability.- How/when to refresh Uber’s server fleet based on useful life, datacenter lease structures, datacenter filling rates and warranty/depreciation expenses.- Amount of resource buffer needed to handle delivery delays, unavailable capacity, and unexpected demand changes.- Amount of datacenter power buffer needed based on likeliness of usage and balanced costs. Supported 2019 Code Yellow (capacity shortage) and 2020-2021 rapid COVID related capacity fluctuations. Developed Infra strategy to front load the purchase of ~7000 servers for a semi-critical refresh, with the optionality of diverting for growth if there were a decrease in COVID cases and increased rides. 6 months later this capacity was diverted and saved the company from another Code Yellow in 2021. Lead TPM driving migrations from 350+ service teams off of bare metal servers in the San Jose region to successfully shut down the region.

 Developing 2 year tactical planning processes and plans for all of Facebook infrastructure, including full decommissioning and cluster turn up schedules based on 'supply vs demand' assumptions and modeling via excel. Training and mentoring the internal capacity planning team and other teams (within Facebook) on processes and standards developed by myself and the Capacity Engineering org. Meticulous rack level planning for all regions within Facebook. Planning rack positions, handling communication for delays, setting priorities for deliveries, managing power, cooling and network at the row/sb/msb/pod/suite/dc/region level, etc. Also working with engineers tool development to manage and drive planning efficiencies for this in-depth level of planning. Working on "Big Bet" cross-org programs that improve scalability, and increase power/network/cooling/planning efficiencies.

 Program Manager for three successful datacenter region turn ups and ~30 individual clusters.  Program Manager of ~15 cluster decommissions, including 2 full datacenters. Primary Lead TPM on the disaster recovery initiative for all of Facebook Infrastructure. Successfully brought down and turned back up multiple datacenter regions in a 24-hour period with no user impact and limited internal issues. Creating processes and documentation for onboarding junior TPMs as well as training/mentoring several TPMs.

 Lead technician for an entire datacenter turn up of 6mW, 750-800 racks. Including power and cooling planning, network turn up and configuration, and server provisioning and maintaining. Led a team of 3-4 Site Ops Technicians to operationally maintain several datacenters. Developed and maintained a full spare parts inventory system within the region.

 Server and network switch installs, including ~50 rack cabling upgrades. Fully organized a datacenter's spare parts and spare servers inventory.