Senior Data Engineer

Designed ETL/ELT pipelines with Airflow, dbt, and AWS Glue to unify HRIS, payroll, and engagement data, building a secure analytics platform for enterprise clients. Partnered with HR and data science teams to support predictive modeling for attrition and workforce planning, enabling leaders to reduce turnover and optimize staffing strategies. Delivered executive dashboards in Tableau and Power BI using AWS Redshift/Snowflake datasets, improving visibility into hiring funnels, diversity metrics, and workforce performance. Established data governance and access controls through AWS IAM and compliance frameworks, ensuring secure handling of sensitive workforce data. Built real-time streaming pipelines with Kafka and Spark on GCP (Pub/Sub, Dataflow, Dataproc, BigQuery), using Scala/PySpark for high-volume processing, powering instant analytics and personalization engines. Developed fraud and anomaly detection pipelines leveraging BigQuery and streaming frameworks, improving platform integrity and preventing financial losses. Designed segmentation workflows and personalization logic that improved targeting of promotions, increasing customer lifetime value by 22%. Led migration of legacy iGaming systems into Google BigQuery and Databricks, reducing infrastructure costs by 30% while scaling to high-volume workloads. Delivered self-serve analytics tools that reduced ad-hoc reporting requests by 40% and accelerated decision-making cycles across HR and gaming clients.

Engineered ML pipelines for Google Ads optimization, leveraging TensorFlow Extended (TFX), Kubeflow, and Apache Beam on Google Cloud to improve ad targeting and bidding. Built a real-time feature store capturing clickstream, keyword, and campaign data, cutting feature duplication and reducing model delivery time by 40%. Designed low-latency Dataflow and BigQuery pipelines to process billions of impressions daily, enabling timely signals for ad-serving models. Collaborated on early LLM-based ad relevance and query understanding prototypes using Transformer architectures and Vertex AI, enhancing semantic matching for search and advertising content. Partnered with research teams to productionize deep learning models that improved click-through rate (CTR) predictions and ad relevance. Implemented monitoring and drift detection systems that safeguarded predictive accuracy across diverse ad markets and geographies. Optimized distributed TPU training workflows, cutting training time by 30% and reducing costs by $1M annually. Developed A/B testing frameworks for ad-ranking models, enabling safe rollouts and rapid rollback of underperforming models. Migrated critical ML workloads to Vertex AI, accelerating deployment velocity and improving standardization across Ads teams. Authored design docs and best practices that unified ML engineering approaches across multiple Google Ads teams.

Built and deployed recommendation algorithms for upsell and cross-sell strategies, increasing average order value by 10%. Designed churn prediction models using gradient boosting and logistic regression, driving retention campaigns that saved millions in recurring revenue. Developed fraud detection models with anomaly detection, reducing false positives in flagged transactions by 18%. Partnered with product teams on A/B and multivariate experiments, applying statistical methods in R to validate outcomes. Created customer segmentation models with clustering methods, enabling personalized marketing that boosted engagement rates. Optimized model training pipelines in Python and Spark on AWS EMR and S3, with components in Scala to improve batch performance and reduce experimentation cycles by 40%. Presented executive-level dashboards and data stories linking predictive insights to business outcomes, shaping roadmap decisions.

Designed and maintained SQL reporting pipelines for claims, billing, and patient records, improving data quality and operational reporting accuracy. Built Tableau/Power BI dashboards to monitor provider performance, patient throughput, and healthcare KPIs, with pipelines supported by AWS Redshift. Partnered with finance teams to detect anomalies in claims data, identifying $2M+ in cost savings and strengthening fraud prevention. Automated recurring reporting workflows in Python (pandas, NumPy), cutting manual preparation by 60% and enabling faster analysis cycles. Developed early predictive models in Python/R to forecast patient demand and optimize clinic staffing. Applied classification techniques on patient outcomes to identify high-risk populations and improve proactive care strategies. Used basic NLP methods to extract insights from clinical notes, informing adherence and treatment effectiveness studies. Supported migration to AWS Redshift, improving query performance and enabling large-scale healthcare analytics adoption.

Contributed to designing and maintaining relational schemas for identity and access management systems, ensuring normalized structures and consistent business rules. Wrote and debugged SQL queries, stored procedures, and triggers that supported reporting and authentication services. Optimized indexing and query execution plans, improving query performance by ~25% in production workloads. Assisted in database migration from Oracle to PostgreSQL, achieving zero data loss and <2 hours downtime. Partnered with senior developers to integrate backend modules with the database layer, reinforcing data consistency and integrity. Reduced recurring reporting errors by 15% by introducing data validation checks and process documentation.