Cloudera has a significantly mature streaming offering on their Data Platform. Data from varied sources such as rich media, text, chat, message queues, etc is brought in to their unified DataFlow platform using Nifi or other ETL/ ELT. After processing these can be directed to one or more of the Op./ App DB, Data Lake (Iceberg), Vector DB post embedding (for AI/ ML), etc.
Streaming in AI/ ML apps help to provide a real-time context that can be leveraged by the apps. Things like feedback mechanism, grounding of outputs, avoiding hallucinations, model evolution, etc all of them require real-time data to be available. So with a better faster data, MLOPs platform Cloudera is looking to improve the quality of the ML apps itself running on them.
Cloudera has also made it easy to get stared with ML with their cloud based Accelarators (AMP). AMPs have support for not just Cloudera built modules, but even those from others like Pinecode, AWS, Hugging Face, etc & the ML community. Apps for Chats, Text summarization, Image analysis, Time series, LLMs, etc are available for use off the shelf. As always, Cloudera continues to offer all deployment options like on-premise, cloud & hybrid as per customer's needs.
