The emergence of open-source large models and the chaining frameworks has made it easier to build AI Systems. The open-source AI infra stack is just being formed. Most observability tools are not open-source and open-data-format. This is where Aim comes in – a developer framework for building end-to-end observability layer for AI Systems.
This talk will cover the infrastructure challenges AI teams face when building AI Systems – from training own models to using LLMs, APIs into production AI Systems.