# Head of System Architecture, Frontier AI Hardware at Sygaldry Technologies

- Company: Sygaldry Technologies
- What the company does: Sygaldry builds quantum-accelerated AI servers to exponentially speed up AI training and inference, cutting compute costs and improving energy efficiency. Backed by Y Combinator.
- Company website: https://www.sygaldry.com
- Type: Startups (AI role)
- Level: Principal and up
- Location: San Francisco
- Work setup: On-site
- Pay: $230K to $400K base salary per year (USD)
- Posted: 2026-04-14
- Apply by: 2026-10-08
- Apply: https://jobs.ashbyhq.com/sygaldry-technologies/3b440563-b6c8-47a0-92d6-575c76cd64eb
- Page: https://www.1752.vc/careers/jobs/sygaldry-technologies-head-of-system-architecture-frontier-ai-hardware/

## About the role

This role is focused on the architecting of high-performing AI hardware, and does not require prior knowledge of quantum computers. This is a hands-on architecture role for someone who enjoys moving between whiteboard-level system design, performance modeling, and lightweight simulation/emulation, with direct influence on hardware, firmware, and software co-design decisions. We will share more detail about this role during in-person discussions.

## What they're looking for

- Architected end-to-end heterogeneous compute systems spanning CPUs, GPUs and novel accelerators
- Analyzed scale-up interconnects (e.g., NVLink, UALink, Infinity Fabric, or equivalent) and their impact on latency, bandwidth, and synchronization of high performance AI hardware
- Designed timing arcs, critical paths, and throughput models across complex multi-device systems
- Identified system-level bottlenecks in training, inference, and reinforcement learning workflows at hyperscaler scale
- Modeled and reasoned about GPU collective operations (e.g., NCCL, RCCL, or equivalents) across nodes, racks, and pods
- Evaluated how collective communication patterns interact with emerging accelerators and non-GPU compute elements

Tags: Technical