正在加载视频...

视频加载失败

DevOps vs LLMOps vs MLOps in 3 mins, let’s go 🚀

16,654 次观看 • 8 个月前 •via X (Twitter)

4 条评论

Practical AI Lab 的头像
Practical AI Lab8 个月前

This framing is useful, but LLMOps feels less like a new ‘ops’ and more like ‘software reliability, except the code is probabilistic’. Evals + traceability are the new tests + logs.

NullPointerMood 的头像
NullPointerMood8 个月前

which one to learn

Stocks Watcher 的头像
Stocks Watcher8 个月前

Great video

Shashank Agarwal 的头像
Shashank Agarwal8 个月前

The ops layer is where most AI projects die. DevOps gives you CI/CD, but LLMOps needs eval pipelines, drift detection, and prompt versioning that didn't exist 2 years ago. We built 68 evaluation scorers at Noveum specifically because "it works in the notebook" means nothing when you're processing 6M+ API calls/mo in production.

相关视频