Meta Superintelligence Labs data infrastructure performance and reliability.

• Responsible for camera storage team’s technical roadmap, project execution, and performance management

• Expanded scope of team to ownership of Cloud Backup product; re-wrote lots of key logic to make backup more flexible, fixed edge cases to improve reliability, and built analytics system to surface more information to customers

• Led series of projects to save >$XXXk in cloud costs by reducing requests to the backend and compressing data structures

• Led initiatives to dramatically improve CI testing pipeline to improve product reliability while increasing team's velocity

• Built new interface on our device to allow applications to request storage space allocations to expand the types of AI features that can run on our devices

• Led project to migrate more than one million storage devices on security cameras to ext4 filesystem, replacing legacy FAT32 filesystem, ultimately reduced customer data lost by two orders of magnitude (C, C++, Bash)

• Designed and built system of Airflow jobs and Python scripts to detect and fix camera storage issues by filtering hundreds of GBs of logs generated by the camera fleet daily, generate signals that measure the real-time health of the fleet, and remotely executing code to fix the storage issue, fixed more than 50,000 cameras with this automated system (Python, SQL)

• Designed and implemented technique for fine-grained performance monitoring of storage device on embedded platform (required Linux kernel modifications, C++ user space bridge process, and backend Python data ingestion pipeline)

• Led effort to improve observability of storage reliability across embedded device fleet (Go job on embedded device and data investigation SQL/Python)

• Planned and implemented C++ tests to exercise several next-gen Hololens hardware components, deployed in production for device manufacturing and thermal validation efforts

• Designed suite of Powershell test scripts to automate and codify process for device failure analysis, working in conjunction with the contract manufacturer and hardware design teams

• Wrote Python scripts to generate reports from factory and environmental validation data to drive understanding of manufacturing process variation in early manufacturing stages