1 day ago
Markham, CanadaSenior
Responsibilities
- Enable and optimize large-scale LLM, VLM, and MoE training on AMD Instinct GPU clusters.
- Build training infrastructure for orchestration, distributed checkpointing, data loading, and storage across multi-thousand-GPU Kubernetes clusters.
- Debug distributed training issues and optimize RCCL collective communication across multi-node topologies.
- Develop monitoring, alerting, compliance, validation, benchmarking, and testing infrastructure for GPU platforms.
- Write and optimize GPU kernels for GEMM, attention, quantized matrix multiplication, and GPTQ/AWQ workloads.
- Enable frontier models on new AMD Instinct GPU silicon and optimize vLLM, SGLang, and TorchServe inference serving.
- Build production quantization pipelines and optimize batching, KV-cache management, speculative decoding, and continuous batching.
- Develop observability and automated analysis tooling for large-scale GPU clusters.
- Collaborate with silicon architecture and pre-silicon teams and contribute to the ROCm ecosystem, SDKs, CI dashboards, documentation, and developer cloud enablement.
Requirements
- Production experience spanning both AI/ML training and inference infrastructure.
- Hands-on experience shipping LLMs on real hardware and optimizing GPU kernels for production performance.
- Experience with AMD Instinct architectures, including MI300X, MI350X, or MI355X, and the RCCL communication library is preferred.
- Experience with anomaly detection, log analysis, observability systems, validation frameworks, proxy benchmarks, or synthetic workload suites for GPU infrastructure is preferred.
- Contributions to PyTorch, vLLM, SGLang, DeepSpeed, or Megatron-LM are preferred.
- Experience with pre-silicon software validation or hardware-software co-verification is preferred.
- Publications or patents in HPC, ML systems, or GPU kernel optimization are preferred.
- Bachelor's, master's, or Ph.D. degree in Computer Engineering, Software Engineering, Computer Science, or a related technical discipline is preferred.
Benefits
- AMD benefits are described through the AMD benefits at a glance resource.
- The position is located in Markham, Canada and is hybrid.
- This role is not eligible for visa sponsorship.
Tech Stack
Categories
About AMD
AMD designs and sells CPUs, GPUs, and adaptive/embedded computing products for PCs, data centers, gaming, and edge devices. Its portfolio includes Ryzen and EPYC processors, Radeon and Instinct graphics, and adaptive SoCs from its Xilinx acquisition, sold to OEMs, cloud providers, and device makers. Founded in 1969 and headquartered in Santa Clara, it is a public company on NASDAQ and supplies semi-custom chips for major game consoles.