Remote GenAI Infrastructure & ML Systems Engineer

Mercor; San Francisco (CA); Remote; USD 140,000 - 210,000; Full time

Job summary

Mercor is seeking engineers to join its cutting-edge GenAI team and help build foundational AI models and LLM infrastructure. You will work across accelerator kernel programming, performance profiling, and high-throughput inference serving.

This fully remote, 40-hour-per-week role emphasizes hands-on ML systems work, with production experience in JAX or PyTorch and a focus on reliability, scalability, and rigorous evaluation of training data for frontier AI systems.

Qualifications

• 2+ years hands-on experience in ML systems, ML infrastructure, or accelerator performance engineering.
• Experience with at least two kernel, profiling, or serving areas listed above.
• Production experience with JAX and/or PyTorch and knowledge of distributed training or compiler work is a plus.
• Ability to engage reliably for 40 hours per week during weekdays.

Responsibilities

• Design tasks across kernels, profiling, debugging and inference serving, including challenging adversarial problems.
• Guide research and engineering teams to improve AI model performance on ML systems and training infra.
• Evaluate ML system tasks and provide clear technical feedback for expert review.
• Develop guidelines and rubrics for kernel optimization, profiler interpretation, and serving throughput/latency.
• Collaborate with subject matter experts to ensure training data consistency.

Skills

Custom kernels

Performance profiling

Distributed workloads

LLM serving at scale

JAX production experience

PyTorch production experience

XLA/StableHLO/MLIR

Tools

JAX

PyTorch

Kineto

Nsight

TensorRT-LLM

Job description

Mercor is seeking engineers to join its cutting-edge GenAI team and help build foundational AI models and LLM infrastructure. You will work across accelerator kernel programming, performance profiling, and high-throughput inference serving.

This fully remote, 40-hour-per-week role emphasizes hands-on ML systems work, with production experience in JAX or PyTorch and a focus on reliability, scalability, and rigorous evaluation of training data for frontier AI systems.