Llm Pre-training & Distributed Engineer (ai Infrastructure)

Hyphen Connect

San Francisco Bay Area, United States
On-site
Deep expertise in 3d parallelism
Experience managing slurm or kubernetes gpu clusters
Strong systems engineering background with c++ cuda python
Hyphen Connect is looking for a highly skilled LLM Pre-training & Distributed Systems Engineer to manage large-scale machine learning training runs and optimize distributed infrastructure. The role requires expertise in GPU clusters and systems engineering, with a focus on orchestrating distributed training using technologies like PyTorch, DeepSpeed, and Megatron-LM

Job Summary

  • This role is essential for orchestrating large-scale machine learning training runs across 1,000+ GPUs using PyTorch, DeepSpeed, or Megatron-LM.
  • The ideal candidate will have a deep understanding of GPU clusters and extensive experience in system engineering to ensure efficient processes.
  • Responsibilities include optimizing networking and memory management while automating checkpointing and failure recovery during month-long training runs.

Matching Summary

Match Score: 85

Hyphen Connect is looking for a highly skilled LLM Pre-training & Distributed Systems Engineer to manage large-scale machine learning training runs and optimize distributed infrastructure. The role requires expertise in GPU clusters and systems engineering, with a focus on orchestrating distributed training using technologies like PyTorch, DeepSpeed, and Megatron-LM.

Skills & Requirements

Must-have

  • Deep expertise in 3D parallelism
  • Experience managing SLURM or Kubernetes GPU clusters
  • Strong systems engineering background with C++ CUDA Python

Nice-to-have

  • Optimizing networking InfiniBand RDMA
  • Automating checkpointing and failure recovery
  • Preventing out-of-memory errors during training

Key Requirements

  • Deep expertise in 3D parallelism (Data, Tensor, Pipeline)
  • Experience managing SLURM or Kubernetes-based GPU clusters
  • Strong systems engineering background (C++, CUDA, Python)

Work Rights

Not specified

Tailored Resume

Cover Letter