– Led a genomic data science project by applying data preprocessing, sample selection, causal inference methodologies, and data visualization to assess patient survival rates and gene expression correlations– Trained Random Forest model with extracted patient features, achieving an accuracy of 91% in predicting patient survival– Identified correlations between gene expressions and patient outcomes in Acute Myeloid Leukemia research by conducting statistical analysis using Python, utilizing Log-rank (Mantel-Cox), Chi-Squared tests, and A/B testing– Evaluated the impact of various prognostic factors on patient outcomes by performing multivariate statistical analysis, utilizing Cox regression techniques, and communicating P-values and survival metrics

– Facilitated operations team's migration from Google Cloud Platform Composer 1 to 2 by engineering Airflow frameworks adhering to SRE practices– Ensured 60% reduction in faulty deployments of data pipelines to GCP

– Processed Terabytes of XML and JSON data daily and stored in Google BigQuery by building scalable data pipelines using Java in Google Dataflow– Devised complex data models for databases with over 3000 features, capturing data relationships through the application of data modeling – Handled data quality issues and ensured optimal performance of data pipelines by designing and implementing dashboards and alerting systems– Reduced external dependency of operations team by automating the deployment process of data pipelines in Google Dataflow – Enhanced efficiency in resource deployment by designing and executing Terraform and Jenkins for CI/CD, Git for version control– Built streaming data pipelines using Beam for transforming data sourced from Teradata, Oracle into Kafka and Google's Pub/Sub services– Automated data loading from HDFS to Hive data warehousing using Talend Studio's Map Reduce and Spark ETL jobs- Conducted comprehensive unit testing, end-to-end testing and performance testing on the data pipelines prior to releasing the pipeline to higher environments– Trained multiple new resources, helping them gain working knowledge on GCP