Data stack for AI driven architecture with Crunchy Data
Crunchy Data's suite of Postgres products support data-driven AI workloads, like generative AI, large language use-cases, semantic search, and predictive analytics. Crunchy Data Warehouse pairs well as a data solution for OpenShift AI.
Not only is AI rapidly changing the world, AI is also rapidly changing. AI has progressed away from the implementation of a monolithic model, and now uses multiple models within a single process. For instance, in a semantic search, a prompt is sent to a classification model, then, executes a search on internal data using nearest neighbor search on embeddings, and finally an LLM summarization step based on original prompt and nearest neighbor.
To find success with AI, organizations need a data stack that enables rapid iterations. The Crunchy Data suite of tools offers benefits like zero-ETL, open-file support, foreign data connections, and easy data movement. Combine those with native vector data storage and indexing to have all the tools necessary for AI success.
Crunchy Data has the open data tools that organizations need:
- Crunchy Postgres: the most advanced open source database for application workloads. Native storage for embeddings, including HNSW indexing.
- Crunchy Data Warehouse: a data lake house with zero-ETL capabilities
Deployments available on cloud, on-prem, and hybrid cloud:
- Crunchy Postgres for Kubernetes: on-prem and hybrid cloud deployments
- Crunchy Bridge: fully-managed public cloud deployments
- Turn-key data solution
When paired with OpenShift AI, at your fingertips, have the full suite of models and data necessary to build a modern AI stack.
- SQL Data Lake House
Use SQL to orchestrate, extract, and analyze data across the organization. Get to solutions faster with the knowledge and toolset already within your organization.
- AI & Vector Optimized
Being the data output from AI classification models, vector data is the backbone of RAG and MCP architecture. Optimize vector data using HNSW indexes to create high-performance queries.
- Automated Deployment
In today’s world, data agility is the difference in days versus months. With a single script, create the same data infrastructure as many times as necessary to support your teams and workloads.
Using Crunchy PostgreSQL for Kubernetes allows us and our customers to order databases in self-service mode, enabling rapid development cycles and substantially shortened deployment times.
David JörgProduct Owner, FOITT