Blog Posts¶
Hands-on walkthroughs published on the AWS blogs, newest first.
2026¶
- OpenFold3 on Amazon SageMaker JumpStart — predict 3D structures of proteins, DNA, RNA, and ligand complexes with OpenFold3-preview on asynchronous Amazon SageMaker AI endpoints, including multi-GPU parallelism, bring-your-own MSAs, and scaling to zero. (13 Nov 2026)
- Distributed training with Ray Train DLCs and EFA on Amazon EKS
— fine-tune Qwen2.5-1.5B with PyTorch FSDP across eight GPUs on two EFA-enabled
g6.12xlargeinstances using the Ray Train DLC and KubeRay, with no image build. (28 Oct 2026) - Multi-node inference with Ray LLM DLCs — serve one LLM across two GPU nodes on Amazon EKS with separate prefill and decode nodes behind a single OpenAI-compatible endpoint, using NIXL KV cache transfer over EFA. (12 Oct 2026)
- Decision models with llama.cpp on Amazon SageMaker AI — deploy a quantized Qwen3.5-4B to a real-time endpoint with the llama.cpp DLC on CPU or GPU, including probabilities over a fixed option set. (12 Oct 2026)
- Real-time voice applications with vLLM-Omni on Amazon SageMaker AI — Part 1 — deploy Qwen3-TTS with the vLLM-Omni DLC and stream generated speech over a persistent bidirectional connection from a Gradio application. (28 Sep 2026)
- Image and video generation with vLLM-Omni on Amazon SageMaker AI — Part 2 — generate an image with FLUX.2-klein through real-time inference, then animate it to video with Wan2.1-VACE through asynchronous inference, from one vLLM-Omni DLC. (28 Sep 2026)
- Speaker-labeled transcription with WhisperX on Amazon SageMaker AI — deploy the WhisperX DLC — Whisper, wav2vec2 forced alignment, and speaker diarization — to real-time and asynchronous endpoints for word-level, speaker-labeled transcription. (24 Sep 2026)
- Simplify and support your TorchServe workloads using Ray Serve DLCs — migrate TorchServe inference workloads to the Ray Serve DLC and deploy a vision-language model on Amazon EKS without maintaining CUDA or framework versions yourself. (9 Sep 2026)
2025¶
- Configure and verify a distributed training cluster with DLCs on Amazon EKS — configure and validate a multi-node training cluster running DLCs on Amazon EKS. (15 Oct 2025)
- Use DLCs with Amazon SageMaker AI managed MLflow — track experiments and manage models with Amazon SageMaker AI managed MLflow from inside a DLC. (18 Sep 2025)
- Deploy LLMs on Amazon EKS using vLLM DLCs — deploy and serve LLMs on Amazon EKS with the vLLM DLC. (14 Aug 2025)
- Fine-tune and deploy Meta Llama 3.2 Vision — fine-tune and deploy Llama 3.2 Vision for generative AI-powered web automation using DLCs, Amazon EKS, and Amazon Bedrock. (29 Jul 2025)
- Streamline deep learning environments with Amazon Q Developer and MCP — set up and troubleshoot DLC environments with Amazon Q Developer and the Model Context Protocol. (22 Jul 2025)
Workshop¶
- LLM deployment on Amazon EKS — deploy and optimize LLMs on Amazon EKS using vLLM DLCs, with sample code.