AWS makes SageMaker AI 'training sessions' generally available, letting users write post-training loops without managing GPUs

Verified (fact)

AWS has added a 'training sessions' feature to SageMaker AI's serverless model customization and made it generally available. Users write their own post-training loops in Python scripts, such as SFT, reinforcement learning, custom loss functions and reward signals, and multi-turn reinforcement learning, while the service handles GPU allocation, distributed execution, failure recovery and checkpointing. AWS introduced three SDK clients and released examples including a GRPO loop on GSM8K and reinforcement learning for competitive programming.

Why it matters (interpretation)

Existing managed fine-tuning often supported only fixed recipes. This feature lets teams run custom reinforcement learning loops such as GRPO without the burden of operating infrastructure. Teams that want to experiment with post-training themselves now have an alternative to Tinker-style services.

Sources

• Official Announcing new ways to customize models with SageMaker AI, without managing GPUs AWS
• Official Write your own training loop — without managing GPUs — with Amazon SageMaker training session APIs AWS

More about AWS

• 2026-10-23 P2 Product Generally available Speech & audio AWS
  Amazon Polly Generative engine gets major upgrade: more expressive and accurate, 14 new voices in 9 locales, Hindi and Finnish GA

  AWS upgraded the Generative engine of Amazon Polly to produce more expressive and accurate speech with low latency. It added 14 new...

• 2026-10-23 P3 Dev tools Generally available Speech & audioMLOps & LLMOpsAgents AWS
  AWS publishes two guides on sentence-level streaming of Kokoro-82M TTS on SageMaker AI and setting autoscaling from measurements

  To reduce response latency for voice agents, AWS described sending each sentence to Kokoro-82M for speech synthesis as soon as the agent...

• 2026-10-23 P3 Product Generally available AI platforms & solutionsAgentsAI security AWS
  AWS publishes guide to deploying AI agents with Amazon Quick in HIPAA-covered environments

  AWS noted that Amazon Quick is a HIPAA-eligible service but that eligibility and compliance are separate, and laid out who is covered by...

• 2026-10-22 P3 Dev tools Generally available AgentsAI platforms & solutions AWSOllama
  AWS shows how to build AAIF, an agent framework that sends simple tasks to local SLMs and complex requests to Bedrock

  AWS introduced the Adaptive Agentic Intelligence Framework (AAIF), built with Strands Agents. Routine tasks are handled by a local small...

• 2026-10-22 P3 Dev tools Generally available Efficiency & infraLanguage models AWSAlibaba
  AWS outlines three ways to train Qwen3 on Trainium with torchtitan

  AWS laid out three ways to train Qwen3 with torchtitan on a single trn2.48xlarge instance. The paths use the Hugging Face Transformers...