How do we optimise the spark job if the base table has 130 billions of records

Question

We are joining multiple tables and doing complex transformations and enrichments. In that the base table will have around 130 billions of records, how can we optimise the sp