High-Performance AI Research Engineer (CUDA/ML)

Metamorphic; Palo Alto (CA); On-site; USD 200,000 - 280,000; Full time

Benefits offered by this job

Visa sponsorship

Competitive compensation

Mentorship and career development

Equity package

Job summary

Metamorphic in Palo Alto seeks Research Engineers to maximize training and inference performance of our foundation models. You will write and optimize GPU kernels, explore quantization, and contribute to efficient, scalable implementations in close collaboration with researchers.

This role blends deep ML research with practical engineering, offering autonomy on a small, high‑impact team and the chance to shape future AI systems at the frontier of neuroscience‑informed AI.

Qualifications

• Bachelor's degree or higher in Computer Science, ML, or related field.
• Strong software engineering skills with track record of building complex systems.
• Proficiency in CUDA, Triton, or similar; experience writing GPU kernels.
• Hands-on experience with mixed-precision and low-precision training; numerical stability tradeoffs.
• Deep knowledge of transformer architectures at implementation level.
• Experience with MoE architectures: routing, load balancing, expert dispatch across GPUs.
• Experience with GPU profiling tools (Nsight Compute, Nsight Systems, PyTorch Profiler).
• Experience integrating high-performance libraries (FlashAttention, cuDNN, Triton, Quack).

Responsibilities

• Write and optimize GPU kernels for training and inference.
• Improve performance via quantization, low-precision training, and MoE routing optimizations.
• Collaborate with researchers to translate theory into scalable implementations.
• Profile code, identify bottlenecks, and drive incremental performance improvements.

Skills

Software engineering

CUDA

Triton

GPU kernels

MoE routing

Profiling tools

Low-precision training

Collaboration

Pair programming

Education

Bachelor's degree or higher in CS/ML

Tools

CUDA

Triton

Nsight Compute/Systems

FlashAttention

cuDNN

Quack

Job description

Metamorphic in Palo Alto seeks Research Engineers to maximize training and inference performance of our foundation models. You will write and optimize GPU kernels, explore quantization, and contribute to efficient, scalable implementations in close collaboration with researchers.

This role blends deep ML research with practical engineering, offering autonomy on a small, high‑impact team and the chance to shape future AI systems at the frontier of neuroscience‑informed AI.