PyTorch MLOps Engineer — Bare-Metal Infra

Second Talent; Singapore; On-site; SGD 103,000 - 156,000; Full time

Job summary

A technology solutions firm is looking for an MLOps Engineer specializing in PyTorch to manage the on-premise infrastructure for advanced training workloads. The role involves architecting training and inference pipelines, ensuring high-quality code, and troubleshooting complex issues. Ideal candidates will have expert knowledge of PyTorch, strong experience in C++ and Python, and a solid background in computer science. This full-time position is based in Singapore and will require a proactive approach to optimizing compute workloads.

Qualifications

• Expert-level knowledge of PyTorch, including DDP, mixed precision training, and TorchScript.
• Advanced programming skills in both C++ and Python.
• Solid background in computer science fundamentals including data structures and algorithms.
• Hands-on experience debugging and tuning bare-metal servers and Linux administration.
• Understanding of low-level networking, distributed training protocols.
• Proven track record of building reliable, reproducible pipelines.

Responsibilities

• Architect, build, and maintain end-to-end training and inference pipelines using PyTorch.
• Develop high-quality tooling in Python and C++ to support model training lifecycle.
• Take ownership of core training codebase for clarity and reproducibility.
• Design workflows for checkpointing, resuming jobs, and model versioning.
• Optimize compute workloads for bare-metal environments.
• Troubleshoot low-level issues, including networking bottlenecks and hardware faults.

Skills

Expert-level knowledge of PyTorch

Advanced programming in C++

Advanced programming in Python

Computer science fundamentals

Debugging bare-metal servers

Low-level networking knowledge

Building reproducible pipelines

Job schedulers experience

Tools

PyTorch

Linux

SLURM

Job description

A technology solutions firm is looking for an MLOps Engineer specializing in PyTorch to manage the on-premise infrastructure for advanced training workloads. The role involves architecting training and inference pipelines, ensuring high-quality code, and troubleshooting complex issues. Ideal candidates will have expert knowledge of PyTorch, strong experience in C++ and Python, and a solid background in computer science. This full-time position is based in Singapore and will require a proactive approach to optimizing compute workloads.