Member of Technical Staff: AI Chip System Modeller

S27a; San Francisco, New York (CA, NY); Hybrid; USD 150.000 - 210.000; Vollzeit

Zusammenfassung

SemiAnalysis LLC is seeking a highly motivated member of the technical staff to join our engineering team to work on system modelling for 100k+ chip AI clusters. This is a rare opportunity to contribute to high-visibility open source projects and shape architecture & model performance for next‑gen AI chips.

If you’re passionate about performance engineering, first principles thinking, and the intersection of hardware and software, you’ll thrive in this role and help drive industry wide impact

Qualifikationen

• Strong Python proficiency and software engineering background.
• Expertise in disaggregated prefill and various parallelism techniques.
• Experience with frontier MoE training and inference workloads.
• Knowledge of GEMM sizes and operator FLOPs in transformers.
• Understanding of Amdahl's law, arithmetic intensity, and scaling.
• Previous ML engineer or kernel programmer experience.

Aufgaben

• Develop a complex Python model to predict performance on AI chips.
• Implement parallelism and inference strategies to build Pareto frontier curves.
• Create microbenchmarks across vendors to calibrate system models.
• Build BoM estimates for performance per total cost of ownership and power.

Kenntnisse

Python

Parallelism

MoE training

GEMM sizes

Amdahl principles

ML engineer experience

Jobbeschreibung

SemiAnalysis LLC is seeking a highly motivated member of the technical staff to join our engineering team to work on system modelling for 100k+ chip AI clusters. This is a rare opportunity to contribute to high-visibility open source projects and shape architecture & model performance for next‑gen AI chips.

If you’re passionate about performance engineering, first principles thinking, and the intersection of hardware and software, you’ll thrive in this role and help drive industry wide impact