Tag
This article provides a comprehensive overview of the complete technology stack for cloud deployment of Transformer inference, covering application scenarios, workload definition, models, inference engines, hardware, observability, and performance optimization, along with future trends.
OpenAI's frontier models and Codex are now generally available on AWS, enabling enterprises to integrate them into existing AWS workflows for faster AI deployment.
Miles Brundage notes that while he struggles to deploy American open weight models on cloud platforms, Chinese models like Kimi and DeepSeek are plug and play.
The Alibaba ATA team deployed Claude Code from a local CLI tool to the cloud, implemented HTTP streaming calls by modifying the SDK, and used sandboxing for multi-user isolation.
Grok CLI now supports plugins and skills, with the Vercel plugin enabling seamless cloud deployment, as shown by generating and hosting a creative coding website.