Hello,
I assume you would be using spark to read data from bronze/silver layer to transform and calculate the summarized/aggregated tables and load it into gold layer.
Additionally, using delta tables especially in gold layer for incremental upserts will allow for efficient processing and updating of data in gold layer tables.
For consumption of data via Power BI, you can also try Databricks SQL which allows for fast response time on large scale data in data lake storage, by leveraging features like delta cache and photon under the hood and is already optimized so no manual configurations or tweaking required.