Worked on the topoteretes/cognee repository to enhance recall precision by introducing explicit support for dataset_ids in the search functionality. This involved updating the API to accept new parameters, prioritizing dataset_ids over dataset names to reduce false positives and improve data retrieval accuracy. Leveraged Python for API development and asynchronous programming, ensuring robust integration with existing systems. Comprehensive unit tests were added to validate the new behavior and safeguard against regressions. The changes addressed recall edge-cases and established a foundation for dataset-centric queries, resulting in more reliable and precise data access for users across the platform during the development period.