Machine Learning Platform Engineer — HPC, GPUs, Cloud

Tower Research Capital; Hong Kong; Hybrid; HKD 941,000 - 1,412,000; Full time

Benefits offered by this job

Generous paid time off policies

Hybrid working opportunities

Free breakfast, lunch and snacks daily

In-office wellness experiences

Company-sponsored sports teams and фит

Volunteer opportunities and charitable

Social events, happy hours, treats

Workshops and continuous learning

Job summary

Tower Research Capital is seeking a strong software engineer to architect and build the next generation of its machine learning research platform. The role emphasizes scalable, observable systems, and close collaboration with quantitative researchers.

You will design pipelines for large-scale experiments, training, and simulation across on-prem HPC and multi-cloud environments, while improving tooling for performance and reproducibility.

Qualifications

• 2+ years designing and building large-scale distributed systems.
• Strong Python programming focusing on clean, high-performance code.
• Experience on Linux-based HPC clusters or cloud platforms.

Responsibilities

• Architect and develop the next generation of Tower’s ML research platform.
• Build infrastructure for large-scale experimentation, training, and simulation.
• Collaborate with researchers to translate workflows into robust platform capabilities.
• Design and optimize distributed training pipelines for GPU workloads.
• Improve experiment management, versioning, and data lineage.

Skills

Python

Distributed systems

Linux HPC

GPU workloads

ML frameworks

Data pipelines

Observability

Tools

Docker

Kubernetes

PyTorch

TensorFlow

JAX

Job description

Tower Research Capital is seeking a strong software engineer to architect and build the next generation of its machine learning research platform. The role emphasizes scalable, observable systems, and close collaboration with quantitative researchers.

You will design pipelines for large-scale experiments, training, and simulation across on-prem HPC and multi-cloud environments, while improving tooling for performance and reproducibility.