From minutes to milliseconds: How CrateDB is tackling AI data infrastructure

The promise of artificial intelligence remains enormous – but a single thing that might hinder it. “The infrastructure that does not maintain Piles AI today is the demands of tomorrow,” leads the last CIO.com article. “CIOS should rethink how to expand the scope of intelligence – not just larger – or in risk of backwardness.”

ATRATEDB agrees – The database company is betting on solving the problem through a “unified data layer for analyzes, research and AI.”

“The challenge is that most information technology systems are adopted, or built, around the pipeline pipeline or an unsafe pipeline, and now you need to reduce the time between production and data consumption,” explains Stefan Castilian, SVP Marketing. “ATRATEDB is very suitable because it can give you an insight into the right data with a large size and the complexity of coordination within a millimeter.”

The Blog Blue Publicity consists of four steps for ATRATEDB is noticed to serve as “the connective tissue between operational data and artificial intelligence systems”; From swallowing, to assembly and insight in an actual time, to data service to artificial intelligence pipelines, to enable the feedback rings between models and data. The speed and diversity of data is the key; Castellani notes reduce inquiries from minutes to millimeters. In manufacturing, remote measurement from machines can be collected in actual time, allowing learning more for predictive maintenance models.

There is another benefit, as Castellani explains. “Some also use ATRATEDB in the factory to get the help of knowledge,” he says. “If a mistake happens, you have a specific error message that appears on your device and says,” I am not an expert in this device, so what does that mean and how can I fix it? ” [you] You can ask an assistant knowledge, which also depends on ARSTB as a vector database, to access information, and withdraw the correct manual instructions appropriate to interact in actual time. “

Artificial intelligence, however, does not stand not for a long time. “We don’t know what [it] Castellani notes within a few months, or even a few weeks. “

Castellani notes excitement about the MCP context protocol, which helps how to provide applications as a context of large language models (LLMS). It is similar to the trend around Apis Enterprise 12 years ago. ATRATEDB McP, which is still in the experimental stage, works as a bridge between artificial intelligence tools and the analysis database. “When we talk about MCP, this approach is largely the same [as APIs] But for LLMS, “explains.

Tech Mahindra is just one of the main partnerships that go ahead in CRATEDB. “We continue to focus on our basics,” Castellani adds. “Performance, the ability to expand … investing our ability to absorb data from more and more data sources, and always minimal[ing] Cumin, whether on swallowing and inquiring. “

Stefan Castilani will talk at AI & Big Data Expo EUROPE on the topic Bring artificial intelligence to data in the actual time-Text2SQL, RAT and TAG with CRATEDBAnd IOT Tech Expo EUROPE on the topic The most intelligent internet operations: wind farm analyzes in real time and the diagnoses driven by artificial intelligence. You can watch the full interview with Stefan below: