Worked on the slick/slick repository to implement advanced UNION support across projections, enabling complex queries that combine different projections of the same table while preserving field order and table identity. Addressed schema construction challenges by introducing a distinctBy helper to remove duplicate fields in union types, ensuring accurate and predictable schema definitions. Improved aggregate query reliability by preserving multi-column order in distinctOn operations, reducing unexpected results in query outputs. The work involved deep AST manipulation and compiler development using Scala and SQL, with a strong emphasis on robust testing to enhance query expressiveness, maintainability, and overall reliability of the codebase.