Helix AI Engineer, Training Performance (CUDA)

Location: San Jose, CA, United States

Salary: $200,000–$400,000 annually

Company

hirify.global is an AI robotics company developing autonomous general-purpose humanoid robots for home and commercial applications.

What you will do

• Optimize distributed training performance for 100B+ parameter models across 100k+ GPUs.
• Influence accelerator selection, cluster topology, scheduling, hardware procurement, and model co-design decisions.
• Write and optimize custom Triton and CUDA kernels, and contribute to kernel compilers such as Triton and Gluon.
• Build monitoring, regression detection, benchmarking, and root-cause analysis tooling for large-scale training jobs.
• Optimize data pipelines, checkpointing, fault tolerance, elastic restart, and model/data parallelism strategies.
• Evaluate AMD, TPU, SRAM-based ASIC, and other emerging accelerators through proof-of-concept ports and benchmarks.

Requirements

• Bachelor’s or Master’s degree in Computer Science, Computer or Electrical Engineering, or a related field.
• 3+ years of AI performance engineering experience, including leadership of large-scale performance improvement projects.
• Deep understanding of GPU architecture, memory bandwidth, compute-bound and memory-bound operations, and occupancy.
• Experience with Nsight Systems/Compute, PyTorch Profiler, HTA, or similar profiling tools.
• Knowledge of NCCL, RDMA, NVLink, InfiniBand/RoCE, and topology-aware placement.
• Strong Python and CUDA/C++ skills, with experience debugging large-scale performance regressions, instability, and hardware-efficiency metrics such as MFU/HFU.

Nice to have

• Experience with heterogeneous or multi-datacenter training and cross-cluster orchestration.
• Open-source contributions to PyTorch, Megatron-LM, vLLM, DeepSpeed, JAX, or related ML systems projects.
• Exposure to AMD GPUs, TPU, Trainium, Inferentia, custom silicon, or heterogeneous fleet management.

Culture & Benefits

• Work on autonomous humanoid robots designed for global deployment.
• Collaborate across AI research, systems engineering, hardware, and infrastructure disciplines.
• Full-time employment with annual base salary of $200,000–$400,000.
• Total compensation may include additional components and benefits depending on the role.