Developed a transcription correction enhancement for the cjpais/handy repository, focusing on robust multi-word phrase matching using Rust. The work introduced an n-gram matching algorithm with greedy selection and length-based filtering to improve the accuracy of custom word corrections while minimizing false positives. Matching logic was refactored into a reusable function, supporting future extensibility and simplifying maintenance. The implementation preserved original punctuation and casing in transcriptions, reducing downstream normalization needs and improving readability. Comprehensive unit tests were created to validate various n-gram scenarios and edge cases, demonstrating strong skills in algorithm design, text processing, and test-driven development throughout the project.