A new version of Genie from Databricks has been released today, focusing on speed and efficiency in AI model training. The update claims to cut training time by up to 50% compared to the previous generation. This follows last year’s launch of the original Genie, which introduced automated model deployment for large-scale data projects. The new Genie introduces a distributed training framework that splits workloads across multiple GPUs without manual configuration. According to Databricks, this reduces the need for specialized infrastructure setups. The company also added real-time performance monitoring tools to track model accuracy during training. Key improvements include support for PyTorch and TensorFlow without additional code changes. Users can now train models on unstructured data like text or images directly. The update also includes better integration with Databricks SQL, allowing seamless transitions between training and querying. Databricks states the changes address common bottlenecks in AI workflows. The company did not disclose pricing for the upgraded Genie but mentioned existing customers would receive updates automatically. Competitors like Snowflake and Google Cloud have also expanded AI tooling in recent months. The release reflects Databricks’ push to dominate enterprise AI infrastructure. The company reported a 35% increase in AI-related revenue last quarter, driven by demand for scalable training solutions.
Source: databricks.com