Worked on the influxdata/iceberg-rust repository to address a bug in the REST catalog’s file IO scheme inference logic. Focused on backend development and cloud storage integration, the solution involved updating the scheme detection process to prioritize metadata_location over warehouse_path, particularly when the warehouse path used an ARN format. This adjustment ensured that S3-backed warehouses were correctly identified, preventing misclassification of the scheme as 'arn' instead of 's3' and improving reliability for downstream data access. The work demonstrated proficiency in Rust and IO scheme parsing, contributing to more robust and maintainable cloud storage operations within the project.