Google gives data scientists one environment with AI functions to simplify their work.
During Big Data London, Google Cloud announced new AI functions intended to make data scientists’ work easier. The improvements are designed to reduce switching between SQL, Python, and visualizations, and to accelerate the building of AI agents, reports SiliconANGLE.
One Environment for SQL, Python, and Visualizations
The improvements are for Colab Enterprise notebooks, which bundle BigQuery and Vertex AI into a single development environment. Native SQL Cells is one such new feature, allowing SQL and Python to be executed side by side. Rich Interactive Visualization Cells automatically convert raw data into interactive graphs. This allows data scientists to perform analyses, queries, and visualizations in the same environment.
The Data Science Agent (DSA) is also getting an upgrade. This Gemini-powered AI assistant can now autonomously build end-to-end analytical pipelines. From data cleaning and exploratory analyses to machine learning training with BigQuery ML and Python-based workflows: the agent supports the entire process.
AI Agents for Unstructured Data
Additionally, Google Cloud is focusing on applications with unstructured data, which are necessary in sectors such as e-commerce and finance. BigQuery continuous queries are getting stateful processing, giving SQL queries “memory” and allowing them to detect patterns in real-time. For example, a sudden spike in credit card transactions can be immediately detected and blocked.
Finally, BigQuery Vector Search is being expanded with automatic and continuous updates of vector databases. Until now, this was a slow process, especially with multimodal data such as images, audio, and video. With the automatic updates, AI agents can now update their memory in real-time.