Tasks:
- * Accelerate and deploy model inference
- * Build self-distillation and self-evolution pipelines
- * Design distributed embodied-model training systems
- * Optimize mixed-precision training and GPU kernels
- * Track experiments and visualize metrics
Perks/Benefits:
Skills/Tech stack required:
[Autoregressive diffusion] [BF16] [C++] [Causal DiT] [CUDA] [Data parallelism] [DeepSpeed] [Diffusion Models] [Distributed Training] [FlashAttention] [FP8] [FSDP] [GPU Performance] [GPU Performance Optimization] [Infiniband] [Linux] [Megatron-LM] [Mixture of Experts] [MoE-DiT] [NCCL] [ONNX Runtime] [Operator fusion] [Performance optimization] [Pipeline parallelism] [Python] [PyTorch] [Quantization] [RDMA] [Tensor Parallelism] [TensorRT] [Transformer] [Triton]
Educational requirements:
N/A
Role(s):
[AI Infrastructure Engineer] [Deep Learning Engineer] [Engineer] [Infrastructure Engineer] [Learning Engineer] [Machine Learning Engineer]