TPU Inference Performance Engineer

Inferact; Singapore; On-site; SGD 200,000 - 400,000; Full time

Benefits offered by this job

Medical, dental, and vision coverage

Equity options

Job summary

Inferact is seeking a TPU performance engineer to enhance vLLM's performance on Google TPUs. You will build and optimize backends, while ensuring benchmarks for correctness and efficiency.

We're looking for candidates with a strong background in computer science, hands-on TPU optimization experience, and familiarity with JAX and XLA. The role is based in Singapore, and offers a competitive salary range of S$200,000 to S$400,000 annually, plus a generous benefits package.

Qualifications

• Hands-on experience building or optimizing TPU workloads.
• Deep understanding of TPU execution and memory behavior.
• Strong performance profiling and benchmarking skills.

Responsibilities

• Build and optimize TPU backends, compiler integrations, and runtime paths.
• Improve production-relevant model serving on TPU.
• Ensure clear correctness, latency, and throughput benchmarks.

Skills

TPU workloads optimization

JAX

XLA

Pallas

Performance profiling

Education

Bachelor's degree in computer science or engineering

Tools

TensorRT-LLM

MLIR

LLVM

Job description

Inferact is seeking a TPU performance engineer to enhance vLLM's performance on Google TPUs. You will build and optimize backends, while ensuring benchmarks for correctness and efficiency.

We're looking for candidates with a strong background in computer science, hands-on TPU optimization experience, and familiarity with JAX and XLA. The role is based in Singapore, and offers a competitive salary range of S$200,000 to S$400,000 annually, plus a generous benefits package.