Production Systems Engineer, AI Systems

About

Meta is seeking a Systems Engineer to join our team working on AI/ML initiatives supporting large-scale AI training and inference. Our servers and data centers are the foundation upon which our rapidly scaling infrastructure operates efficiently to deliver our innovative services. The Production Systems Engineering team is responsible for the end-to-end hardware lifecycle of all Meta servers, including prototyping of experimental HW, pre-production hands-on system and hardware debugging and stress testing, enabling production-ready system monitoring, automated provisioning and automated remediation of issues, ultimately certifying new platforms for mass production for AI at datacenter scale. Production Systems Engineers have a large swath of cross-functional partners they work closely with, for example, HW/SW co-design teams, hardware designers, networking teams, system manufacturers, component vendors, capacity engineering, production engineering, production services, and data center operations teams, to enable new systems that will be deployed in our production data centers. We are looking for a candidate who can support the scale up and scale out of network technologies for Meta AI systems that are powering Meta's tremendous leaps in the AI space. A strong fit for this role includes someone is knowledgeable about network technologies (NICs, Switches, Optics, DACs, Protocols-TCP/IP, RDMA) and has hands-on experience supporting them through at least a couple of hardware/software (firmware, driver) lifecycle phases: design and bring-up, server integration, system validation, supporting customer deployment, production issue triage, rolling out new features in FW/Driver.

Responsibilities

• Lead integration of scale up (e.g. NVlink, XGMI, RoCE) and scale out (e.g. NICs) interfaces for AI Platforms
• Develop understanding of Collective Communication patterns/ AI workloads and incorporate as part of new product introduction (NPI)
• Proactively create experiments and tooling to detect, reproduce and diagnose hardware/firmware/software issues
• Contribute to enabling hacks for future technology explorations in AI space
• Troubleshoot, diagnose and root-cause system failures and isolate the components/failure scenarios while working with internal & external partners
• Develop visibility through data visualization and implement systemic solutions to hardware health issues
• Leverage production experience to drive external and internal teams to continuously improve product quality

Minimum Qualifications

• Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
• 8+ years of work experience in one or more domains such as: Network ASIC/Platform Development (Silicon/Switch Platform design or bring-up or characterization), Network Product Deployment and Customer Support (Switches, NICs), Interconnect Technologies (e.g. Optics, DAC)
• Knowledge of TCP/IP and experience using tools like iperf/uperf
• Knowledge of server architecture and components
• Experience working with Linux
• Hands-on troubleshooting and debug experience Experience working with RDMA/RoCE, including scale-out networks
• Experience with Python scripting
• Experience working with Network Interface Cards (NICs)
• Experience working with AI server systems
• Experience working with large scale deployments