Remote Deep Learning Engineer – Distributed Training

BairesDev; Mexico; On-site; PHP 7,299,000 - 10,949,000; Full time

Benefits offered by this job

Remote work 100%

Competitive compensation (USD or local

Home office setup provided

Flexible hours

Paid parental leave & holidays

Global, multicultural team

Job summary

BairesDev is seeking a Deep Learning Engineer (Distributed) to lead the development and optimization of large-scale model training across multi-GPU and multi-node environments. You will maximize throughput and scale complex models on high-performance compute clusters.

Join a global, fully remote team and collaborate with researchers and systems experts to push the boundaries of distributed ML, using PyTorch, DeepSpeed, Triton, and NVIDIA tools to deliver efficient training and inference at scale.

Qualifications

• 4+ years of experience in ML engineering or distributed systems.
• Expertise in PyTorch and distributed training (DDP, FSDP, DeepSpeed).
• Proficient with CUDA and Nvidia profiling tools (Nsight).
• Experience with Triton or TensorRT for performance optimization.
• Excellent English communication.

Responsibilities

• Design and implement distributed training pipelines using PyTorch DDP, FSDP, and DeepSpeed.
• Optimize model performance through mixed-precision training and sharding techniques.
• Identify and resolve system bottlenecks with Nvidia Nsight.
• Develop and deploy optimized model kernels using Triton or TensorRT.
• Collaborate with research and systems teams on distributed state and checkpointing.

Skills

PyTorch

Distributed training

DeepSpeed

DDP

FSDP

Triton

TensorRT

CUDA

Nsight

English

Job description

BairesDev is seeking a Deep Learning Engineer (Distributed) to lead the development and optimization of large-scale model training across multi-GPU and multi-node environments. You will maximize throughput and scale complex models on high-performance compute clusters.

Join a global, fully remote team and collaborate with researchers and systems experts to push the boundaries of distributed ML, using PyTorch, DeepSpeed, Triton, and NVIDIA tools to deliver efficient training and inference at scale.