Deepak Yadav contributed to the snowflakedb/snowpark-python and xupefei/spark repositories, focusing on enhancing data processing and API compatibility. He implemented Spark API compatibility for apply_in_pandas, addressing Spark-specific data structures and column naming, and expanded pivot functionality to support multi-aggregate operations with GROUPBY, using JOIN-based composition for Spark-like output. Deepak also improved schema inference robustness by adding error handling and fallback mechanisms for JSON, ORC, and AVRO formats, and enabled user-provided schemas for Parquet reads. His work, primarily in Python and SQL, emphasized reliable data engineering, thorough testing, and improved metadata integrity for resilient data pipelines.