Highlights:

  • Verta promises to add more efficient retrieval-augmented generation (RAG) features to neural networks. With the use of RAG, a machine learning technique, an AI model’s knowledge base can be increased without requiring it to be retrained.
  • Verta’s technology will enhance Cloudera’s machine-learning capabilities as part of its flagship data platform. These features simplify tasks like deploying and training AI models in production.

Cloudera Inc. recently announced the acquisition of Verta Inc., a venture-backed company that creates tools for creating AI applications.

The Cloudera Verta agreement’s terms were not made public. The acquisition occurred approximately three years after two private equity firms paid USD 5.3 billion to acquire Cloudera, a significant data management and analysis software provider.

Charles Sansbury, CEO of Cloudera, stated, “Cloudera is acquiring Verta’s AI Operational platform to strengthen our team and accelerate our operational AI capabilities. We are moving quickly to provide our customers with the technologies they need to drive their data and AI initiatives, and we have no plans to slow down.”

Before the Cloudera Verta agreement, Verta raised over USD 15 million in funding from investors, including Intel Capital. The software platform bearing its name offers pre-packaged artificial intelligence models, datasets for training said models, and example prompts. According to the company, these starter kits will expedite AI development projects by eliminating the requirement for software teams to start from scratch on every project.

Verta also promises to add more efficient retrieval-augmented generation (RAG) features to neural networks. With the use of RAG, a machine learning technique, an AI model’s knowledge base can be increased without requiring it to be retrained. This lowers costs by drastically reducing the need for infrastructure.

Putting RAG into practice by hand can take a lot of work. Verta claims its platform can reduce the workflow time to a few minutes. Additionally, developers are spared from writing original code.

Enhancing a neural network’s output quality can be achieved in more ways than just giving it access to more data through RAG. Occasionally, minor textual changes to a prompt can also improve AI responses. However, pinpointing the precise changes that result in quality improvement is frequently an uphill task.

PromptBrew, a tool on Verta’s platform, automates the process. The company claims that developers can use the tool to enter a prompt and have it generate different versions of the input text, which will result in higher caliber AI responses. Software teams can enhance the output of their AI applications for end users by utilizing PromptBrew.

Verta’s technology will enhance Cloudera’s machine-learning capabilities as part of its flagship data platform. These features simplify tasks like deploying and training AI models in production. Cloudera also provides pre-packaged machine learning building blocks, or “AMPs,” which are software bundles akin to Verta’s AI starter kits to save developers’ time.