Remote Senior HPC & GPU Architect

Bright Vision Technologies; Alpharetta (GA); Remote; USD 130,000 - 180,000; Full time

Job summary

Bright Vision Technologies is seeking a highly experienced Parallel Computing Engineer to optimize AI, ML, and scientific workloads on enterprise‑grade GPU architectures.

The ideal candidate will have 10+ years in HPC, CUDA, and distributed computing, with a track record of designing scalable GPU solutions for enterprise and research environments, including experience with NCCL, MPI, and CUDA-aware frameworks.

Qualifications

• 10+ years of professional GPU programming or HPC experience.
• Expert-level CUDA C/C++ and GPU architecture knowledge.
• Experience with NCCL, MPI, and distributed GPU frameworks.
• Experience integrating custom GPU kernels into ML frameworks (PyTorch, TensorFlow, JAX).
• Strong C/C++ debugging and performance optimization skills.
• Experience with cloud platforms or large GPU clusters.

Responsibilities

• Design and optimize CUDA kernels for AI, DL, and scientific computing.
• Profile and optimize GPU workloads using Nsight tools.
• Optimize memory usage, kernel performance, and multi-GPU scaling.
• Build distributed training/inference architectures using NCCL, MPI.
• Develop custom GPU operators for PyTorch, TensorFlow, JAX, or Triton.
• Improve training and inference for LLMs and AI workloads.
• Collaborate with AI researchers and software architects.
• Create benchmarking, regression testing, and optimization pipelines.
• Evaluate new GPU technologies and accelerator architectures.
• Mentor engineers in GPU optimization and HPC best practices.

Skills

CUDA C/C++

GPU architecture

HPC

Parallel programming

Performance analysis

ML frameworks integration

Debugging & profiling

Distributed GPU computing

Cloud platforms

Communication libraries

Education

Bachelor's or Master's in CS/CE/EE

Tools

NCCL

MPI

CUDA-aware MPI

NVIDIA Nsight

Job description

Bright Vision Technologies is seeking a highly experienced Parallel Computing Engineer to optimize AI, ML, and scientific workloads on enterprise‑grade GPU architectures.

The ideal candidate will have 10+ years in HPC, CUDA, and distributed computing, with a track record of designing scalable GPU solutions for enterprise and research environments, including experience with NCCL, MPI, and CUDA-aware frameworks.