In-Database ML / AI Layer, Evaluation

Overview

A local deployment of MindsDB, the open-source platform that brings machine-learning and LLM capabilities into databases via SQL, stood up for evaluation. The folder is a Docker Compose setup pulling the official mindsdb/mindsdb image plus its data volume.

Why It Exists

To evaluate MindsDB’s “AI tables” approach: training/querying ML and LLM models directly through SQL, which lowers the barrier to predictive and generative features for data-centric applications.

What We Built

No custom code, a Docker Compose deployment of the upstream image (mindsdb/mindsdb) with a local data directory, used to run and explore the platform hands-on.

Technologies & Approach

Docker Compose running MindsDB, which exposes ML/LLM functionality through a SQL interface over connected databases. Reviewed for how it abstracts model training and inference as SQL.

Outcome / Impact

Gave the studio practical exposure to in-database AI/ML and SQL-native model serving, informing options for embedding predictive/LLM features into data workflows. Documented as evaluation/R&D.

Capabilities Demonstrated

• Evaluating in-database ML / SQL-native AI platforms
• Hands-on deployment of containerized AI infrastructure