Tasks:
- * Analyze and optimize data loading and communication efficiency
- * Build large-model inference frameworks
- * Deploy large models to production
- * Develop distributed training platforms
- * Develop model training and inference strategies
- * Develop performance monitoring and profiling tools
- * Implement context caching
- * Implement model quantization
- * Implement prefill decode disaggregation
- * Improve hardware utilization
- * Integrate and optimize high-performance computing technologies
- * Maintain and extend distributed training frameworks
- * Optimize inference speed memory usage and energy consumption
- * Optimize training efficiency and resource utilization
- * Support online and offline inference
Perks/Benefits:
Skills/Tech stack required:
[C++] [CUDA] [CUDNN] [DeepSpeed] [Distributed Training] [Docker] [GPU] [High Performance] [High-Performance Computing] [Horovod] [JAX] [Kubernetes] [Language Models] [Large Language Models] [Mixed Precision] [Mixed-precision training] [Model Inference] [Model Pruning] [Model Quantization] [MPI] [NCCL] [NPU] [Performance Computing] [Performance optimization] [Python] [PyTorch] [Ray] [TensorFlow] [Terraform] [Transformer Models]
Educational requirements:
[Bachelor's Degree]
Role(s):
[Distributed Training Engineer] [Engineer] [Inference Infrastructure Engineer] [Infrastructure Engineer] [Machine Learning Infrastructure Engineer] [Training Engineer]