Himanshu Maheshwari

Developed ELT pipelines on GCP BigQuery integrating AML AI for high-risk transaction detection with explainable outputs. Processed 50M+ transaction records for AML surveillance using BigQuery, Jupyter Notebook/Workstation for transformations, and reloading data back to BigQuery. Automated ingestion, model training, and risk scoring workflows improved detection accuracy by 20% and reduced manual intervention by 20+ hours/month. Collaborated with stakeholders for validation and deployment of AML AI use cases. Reduced BigQuery costs by 25% through query tuning, partition pruning, and storage optimization. Delivered full AML lineage and reproducibility framework for audit readiness and improved pipeline SLA compliance by 35%.

Developed ELT pipelines on GCP BigQuery integrating AML AI for high-risk transaction detection with explainable outputs. Designed AI-supported ELT pipelines processing 50M+ transaction records. Used Tangerine customer and transaction data from BigQuery as the source; performed ELT transformations via Jupyter Notebook/Workstation, reloading the transformed data back into BigQuery. Automated ingestion, model training, and risk-scoring workflows improved detection accuracy and reduced manual intervention by 20+ hours per month. Collaborated with business stakeholders for validation, fine-tuning, and deployment of AML AI use cases. Reduced BigQuery job costs by 25% through query tuning, partition pruning, and storage optimization.

Developed ELT pipelines on GCP BigQuery using DBT for AML use cases at Scotiabank, processing 50M+ transactions. Automated ingestion, model training, and risk-scoring workflows; improved AML detection accuracy by 20% and reduced manual intervention by 20+ hours/month. Optimized queries with partitioning and storage improvements, achieving 25% cost reduction. Delivered AML lineage and reproducibility framework to support audit readiness and improved pipeline SLA compliance by 35%.