Kevin Jiao contributed to the apache/iceberg-python repository by delivering a targeted bug fix that addressed column projection handling for partitioned Iceberg tables during schema evolution. Using Python and data engineering skills, Kevin ensured that the full table schema is consistently passed through ArrowScan and into partition type construction, preventing read-time errors when projecting columns that exclude partition fields. He implemented an end-to-end regression test to validate the solution across various schema evolution and projection scenarios. This work improved the stability and reliability of data reads for partitioned tables, reducing production incidents and supporting safer analytics over evolving datasets.