Remote ML Infrastructure Engineer: GPU Training Platform

Bright Vision Technologies; United States; Teletrabalho; USD 100 000 - 150 000; Tempo integral

Resumo da oferta

Bright Vision Technologies seeks an experienced ML Infrastructure Engineer to design, build, and operate the platform layer powering large-scale AI training and inference workloads. You will work with GPU clusters, scheduler and storage systems to ensure reliability, efficiency, and cost control across hybrid configurations.

The role focuses on integrating PyTorch, DeepSpeed, JAX, and other tools, while improving developer experience for ML researchers.

Qualificações

• Bachelor’s or Master’s degree in Computer Science or related field.
• Six+ years of infrastructure, platform, or HPC engineering experience.
• Hands-on experience operating GPU clusters or large-scale ML training infrastructure.
• Strong proficiency in Python and at least one systems language (Go or C++).
• Deep understanding of distributed training, accelerator architectures, and collective communication.
• Experience with Kubernetes, Slurm, Ray, or similar ML scheduling systems.
• Strong Linux, networking, and high-performance storage knowledge.
• Experience with major cloud ML infrastructure offerings.
• Strong software engineering practices including testing, CI/CD, and code reviews.
• Excellent communication and cross-functional collaboration skills.

Responsabilidades

• Design and operate GPU and accelerator infrastructure for training and inference across on-prem, cloud, and hybrid setups.
• Build scheduling, queuing, and resource-sharing systems to maximize accelerator utilization.
• Integrate frameworks like PyTorch, JAX, DeepSpeed, and Ray Train into a unified platform.
• Operate high-performance storage systems and data pipelines for training data at near-line-rate.
• Design networking architectures supporting RDMA, InfiniBand, NCCL, and high-bandwidth communication.
• Implement observability for AI workloads including utilization and failure analytics.
• Develop tooling to enable researchers to launch experiments safely and efficiently.
• Drive cost optimization across compute, storage, and networking through scheduling and spot capacity.

Conhecimentos

Python proficiency

Distributed training concepts

Linux proficiency

Cross-functional collaboration

Formação académica

Bachelor’s or Master’s degree in Computer Science

Ferramentas

GPU clusters

Kubernetes

Slurm

Ray

PyTorch

DeepSpeed

FSDP

Megatron-LM

NCCL

Descrição da oferta de emprego

Bright Vision Technologies seeks an experienced ML Infrastructure Engineer to design, build, and operate the platform layer powering large-scale AI training and inference workloads. You will work with GPU clusters, scheduler and storage systems to ensure reliability, efficiency, and cost control across hybrid configurations.

The role focuses on integrating PyTorch, DeepSpeed, JAX, and other tools, while improving developer experience for ML researchers.