AI Systems Performance & Simulation Architect

Detroit Venture Partners, LLC.; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Benefits offered by this job

Visa sponsorship

Competitive salary

Equity

Health coverage

Offsite events

Unlimited PTO

Job summary

Sygaldry Technologies seeks a researcher to model and simulate high‑performing AI hardware. This role focuses on performance modeling and simulators for large‑scale distributed ML systems and AI acceleration platforms, interfacing with ML researchers and system architects.

You will analyze workload parallelism, data movement, and network bottlenecks, explore hardware/software tradeoffs, and help guide next‑generation AI system design through rigorous analysis and fast design‑space exploration.

Qualifications

• PhD in Computer Engineering or Masters with relevant research background.
• Strong research track record with publications at top venues (e.g., ISCA, MICRO, MLSys, SOSP).
• Foundations in ML systems, distributed computing, and performance modeling.
• Experience with large-scale ML training/inference and distributed execution frameworks.
• Understanding model/data parallelism, collective communication, and distributed scheduling.
• Experience building simulators or system-level benchmarking tools.
• Familiarity with GPU clusters, high-performance networking, and AI accelerators.

Responsibilities

• Develop performance models and simulators for large-scale distributed ML systems and AI acceleration platforms.
• Model end-to-end training and inference performance across compute, memory, networking, and communication.
• Analyze workload parallelism, scheduling, and data movement to identify bottlenecks.
• Explore system architectures and hardware/software tradeoffs to improve performance and scalability.
• Build simulation frameworks for rapid design-space exploration across distributed GPUs and accelerators.
• Collaborate with ML researchers, compiler engineers, and system architects to guide AI system design.

Skills

ML systems

Distributed computing

Performance modeling

GPU clusters

AI accelerators

Simulators

Education

PhD in Computer Engineering

Masters in Computer Engineering

Tools

CUDA

MPI

Horovod

Job description

Sygaldry Technologies seeks a researcher to model and simulate high‑performing AI hardware. This role focuses on performance modeling and simulators for large‑scale distributed ML systems and AI acceleration platforms, interfacing with ML researchers and system architects.

You will analyze workload parallelism, data movement, and network bottlenecks, explore hardware/software tradeoffs, and help guide next‑generation AI system design through rigorous analysis and fast design‑space exploration.