Sr. Systems Architecture Engineer, Megapod

What to Expect

The Tesla Energy Products team is building Megapod to deliver high-density AI training and inference compute at grid scale, designed for rapid deployment. As a Systems Architect, you are the design leadership that defines Megapod compute and networking requirements and architecture by leveraging your physics fundamentals, systems optimization thinking, and simulation tools. You will thrive in a high-ambiguity, high-speed environment and drive architecture decisions for a complex hardware product with limited precedent. This role involves close collaboration with design, engineering, manufacturing, and product teams, and the ability to distill highly technical information and communicate it clearly to other teams and management. Our team is small, committed, and fast-moving. If this sounds like you, Tesla just might be the place to make your mark.

What You’ll Do

• Define Megapod training and inference compute requirements for various GPU and compute tray architectures, covering power, thermal density, voltage rails, form factor, and interconnect
• Lead compute and networking trade studies on density, interconnect, network power, redundancy, cooling, layout, and cost, and deliver recommendations that support overall system architecture decisions
• Build and use models of training and inference load profiles to size compute power, cooling demand, and network capacity. Validate models against test data
• Define power, cooling, and mechanical interface requirements driven by compute and networking architecture
• In coordination with the reliability team, set compute and networking reliability and availability requirements, including redundancy and service strategy
• Facilitate Megapod networking architecture and global compliance requirements definition and documentation by collaborating with internal and external customers
• Support tray and rack compute and networking interconnect architecture, including connectors, cables, optics, I/O placement, and service access, and manage their impact to overall Megapod product architecture and performance
• Define site computing and networking interface requirements for Megapod deployment

What You’ll Bring

• Systems-level thinker with evidence of driving architecture decisions across products, disciplines, and teams
• Fundamentals in compute architecture and networking infrastructure, with working knowledge of GPU architectures for training and inference, including power, thermal density, memory, interconnect, and form factor
• Experience with tray and rack architecture, including power and thermal envelopes, interconnect, I/O placement, and service access
• Ability to generate and interpret models of training and inference load, power, cooling, and network capacity, and integrate recommendations into product architecture
• Experience with high-performance or datacenter networking, including fabrics, topology, optics, cables, I/O density, and network power
• Fluency in Python/MATLAB
• Specification definition (performance, functional, dimensional, manufacturing, compliance)
• Detail-oriented personality with effective documentation and reporting skills
• Degree in Electrical Engineering, Computer Engineering, Physics, or equivalent experience