Business analysts know what question they want answered but not how to write the SQL query to get it. Our natural language query builder lets them type "show me monthly revenue by product category for the last 6 months, compared to the same period last year" and get a working query against their data warehouse — complete with the right joins, date calculations, and comparison logic. You'll build the query generation pipeline: parsing natural language questions into structured intent representations, mapping business terms to database schemas using client-specific glossaries, generating syntactically correct SQL for different warehouse dialects, and validating results against expected output shapes. The hardest part isn't generating SQL — it's handling the ambiguity in how business users describe what they want. The analytics tools team is four engineers. You'll focus on the NL-to-SQL translation layer, starting with simple aggregation queries and building toward complex multi-table joins, window functions, and temporal comparisons.