GPU Performance Engineer – AI Inference & Systems

gimlet; San Francisco (CA); Sur place; USD 180 000 - 280 000; Plein temps

Résumé du poste

Gimlet in San Francisco is hiring a Member of Technical Staff to characterize and improve the performance of production AI inference across our heterogeneous accelerator fleet. You will study kernel execution, compiler output, runtime behavior, and data movement to reduce latency and increase throughput.

You will build measurement systems and performance methodologies to identify bottlenecks, validate improvements, and guide changes across kernels, runtimes, and distributed execution in

Qualifications

• Experience building performance measurement plans and benchmarks.
• Ability to translate measurements into production optimizations.
• Bachelor's degree or equivalent in a relevant field.

Responsabilités

• Characterize and improve production AI inference performance across heterogeneous accelerators.
• Analyze kernel execution, memory movement, and scheduling to reduce latency.
• Develop measurement systems and performance models to guide optimizations.
• Collaborate with kernel, compiler, ML systems, and distributed engineering teams.

Connaissances

Software engineering fundamentals

Profiling GPU/accelerator workloads

GPU execution/memory hierarchies/sync

Design controlled experiments

Formation

Bachelor's degree in a relevant field

Outils

Nsight Systems

Nsight Compute

rocprof

Omniperf

VTune

Perfetto

CUDA

Triton

Description du poste

Gimlet in San Francisco is hiring a Member of Technical Staff to characterize and improve the performance of production AI inference across our heterogeneous accelerator fleet. You will study kernel execution, compiler output, runtime behavior, and data movement to reduce latency and increase throughput.

You will build measurement systems and performance methodologies to identify bottlenecks, validate improvements, and guide changes across kernels, runtimes, and distributed execution in