Senior AI Systems Architect: Scalable Distributed Training

Cerence AI; United States; On-site; USD 180,000 - 240,000; Full time

Job summary

Cerence Inc. is seeking an experienced ML Systems Engineer to design and run distributed training systems for large neural networks across GPU clusters.

You will optimize multi-node, multi-GPU execution to maximize throughput and tenacity, while diagnosing bottlenecks across compute, memory, and network. Responsibilities include productionising large‑model training pipelines with PyTorch Distributed, Megatron‑LM and DeepSpeed, and improving training stability at scale.

Qualifications

• Deep hands-on experience with distributed systems or ML systems.
• Experience running large-scale workloads on GPU clusters.
• Production experience with PyTorch distributed training.
• Strong understanding of data, tensor, and pipeline parallelism.
• Low-level understanding of GPU communication and networking.

Responsibilities

• Design and operate distributed training systems for large neural networks across GPU clusters.
• Optimise multi-node, multi-GPU execution to maximise throughput and utilisation.
• Diagnose and resolve bottlenecks across compute, memory, and network.
• Improve training stability and fault tolerance at scale.
• Partner with research and applied ML teams to productionise training pipelines.

Skills

Distributed systems experience

Large‑scale GPU workloads

PyTorch distributed training

Parallelism knowledge (data/tensor/pip

GPU communication & networking

GPU orchestration (Slurm/Kubernetes/R​

NCCL/RDMA/InfiniBand/NVLink

Megatron‑LM/DeepSpeed

Activation checkpointing/ZeRO offload

Tools

Slurm

Kubernetes

Ray

RunAI

NCCL

RDMA

InfiniBand

NVLink

Job description

Cerence Inc. is seeking an experienced ML Systems Engineer to design and run distributed training systems for large neural networks across GPU clusters.

You will optimize multi-node, multi-GPU execution to maximize throughput and tenacity, while diagnosing bottlenecks across compute, memory, and network. Responsibilities include productionising large‑model training pipelines with PyTorch Distributed, Megatron‑LM and DeepSpeed, and improving training stability at scale.