TL;DR

Performance Engineer (AI): Architecting and implementing foundational GPU performance systems to power large language models with an accent on maximizing GPU utilization, custom kernel development, and distributed system architectures. Focus on designing and optimizing GPU kernels, orchestrating multi-node GPU clusters, and improving inference efficiency at scale.

Location: Hybrid with at least 25% in-office presence in San Francisco

Salary: $315,000 - $560,000 USD annually

Company

%hirify_global% is a public benefit corporation focused on creating reliable, interpretable, and steerable AI systems with a collaborative and impact-driven research culture.

What you will do

Architect and implement GPU performance optimizations for large language models
Develop custom kernels and optimize tensor core usage
Design distributed communication strategies for multi-node GPU clusters
Optimize training and inference pipelines for AI models
Build performance modeling frameworks to predict and improve GPU utilization
Collaborate with hardware vendors to influence future accelerator capabilities

Requirements

Must have at least a Bachelor's degree or equivalent experience
Hybrid work format with minimum 25% office presence
Visa sponsorship available but not guaranteed for all candidates
Deep experience with GPU programming and optimization at scale
Strong knowledge of CUDA, Triton, CUTLASS, and ML frameworks like PyTorch and JAX
Experience with distributed systems and performance engineering

Nice to have

Experience with low-precision quantization techniques (INT8/FP8)
Familiarity with kernel fusion and memory bandwidth optimization
Background in large-scale training infrastructure and fault tolerance

Culture & Benefits

Competitive compensation including equity and benefits
Generous vacation and parental leave
Flexible working hours
Collaborative office environment in San Francisco
Commitment to diversity and inclusion