Paragon Alpha are working with a leading AI research organisation developing next-generation foundation models and are looking for an experienced Infrastructure Engineer to help scale the infrastructure behind frontier AI.
In this role, you’ll design and optimise large-scale distributed training systems, working closely with researchers to build production-ready infrastructure capable of training models across thousands of GPUs. You’ll focus on improving performance, stability and efficiency while enabling rapid experimentation and iteration.
Stack: Kubernetes, GPUs, Megatron/DeepSpeed
We’re looking for engineers with experience building distributed training systems for large machine learning models, strong knowledge of frameworks such as Megatron or DeepSpeed, expertise in model parallelism, GPU optimisation and NCCL, and a track record of solving complex performance challenges in large-scale ML environments.
This is an opportunity to join an OpenAI or Anthropic competitor at an early stage, with great opportunities for thus equity and impact on the firms growth and direction.
…