High-Performance ML Backend Engineer

United States Digital Space LLC; Singapore; On-site; SGD 70,000 - 120,000; Full time

Job summary

the company in Singapore is seeking a Backend performance engineer focused on optimizing the company's mid-platform recommendation models. You will lead kernel library design, implement graph and code-generation optimizations, and work with AI/algorithm teams to diagnose bottlenecks.

The role emphasizes strong C/C++ skills, experience with model compilation stacks, and deep learning frameworks; onboarding expected by late 2027 with opportunities for impactful research.

Qualifications

• Bachelor's or master's degree in computing or related discipline.
• Familiar with model compilation stacks (TVM, MLIR, XLA) and experience in development/optimization.
• Proficient in C/C++ with understanding of CPU/GPU architecture and high-performance kernel development.
• Familiar with deep learning frameworks (TensorFlow, PyTorch, OneFlow) and their computation graphs.
• Strong independent thinking, problem decomposition, and ability to troubleshoot performance bottlenecks.

Responsibilities

• Responsible for computational performance optimization of the company's recommendation mid-platform models; tune for inference/training bottlenecks.
• Lead design and development of high-performance kernel libraries and general-purpose/custom kernels; ensure peak kernel performance.
• Develop model compilation optimization tech across graph optimization, kernel fusion, and code generation; enhance mid-platform compilation system.
• Collaborate with business and algorithm teams to diagnose performance issues and provide full-stack optimization solutions.

Skills

C/C++ development

Model compilation stacks

DL frameworks

Independence & problem solving

Bachelor's/Master's in computing

Education

Bachelor's/Master's degree in computing or related discipline

Job description

the company in Singapore is seeking a Backend performance engineer focused on optimizing the company's mid-platform recommendation models. You will lead kernel library design, implement graph and code-generation optimizations, and work with AI/algorithm teams to diagnose bottlenecks.

The role emphasizes strong C/C++ skills, experience with model compilation stacks, and deep learning frameworks; onboarding expected by late 2027 with opportunities for impactful research.