giridhar reddy tatiparthi

EBS – Fintech | FDLH & FCP Teams | Hybrid GCP + Azure • Engineered and maintained production-grade Spark-based ETL pipelines on Google Dataproc, processing daily store and eCommerce sales transaction data from upstream SAS systems landed in GCS, directly supporting finance reporting and reconciliation workflows. • Designed partition-based incremental processing using batch_id and batch_date driven ingestion logic with rolling 7-day timestamp windows, storing transformed outputs in Parquet format partitioned by business date — reducing full dataset scans and improving query performance for downstream consumers. • Built and maintained a layered data processing architecture (Raw → Curated → Enriched → Consumption), applying business transformation rules across FDLH and FCP aligned to downstream finance requirements. • Supported hybrid cloud data architecture (GCP + Azure), enabling downstream consumption through Azure Data Factory (ADF) and Azure Databricks, while operating within Automic for SLA-driven batch orchestration and touch file-based data arrival validation. • Proactively raised and tracked Jira tickets for pipeline failures, job execution errors, data quality issues, schema changes, and SLA breaches — coordinating resolution across upstream SAS teams and downstream reporting stakeholders to minimize business impact. • Performed data validation and reconciliation including record count verification, data mismatch analysis, and root cause investigation, ensuring data accuracy across finance-critical reporting pipelines. • Monitored end-to-end production pipelines supporting the BigQuery-based consumption layer, triaged and resolved job failures, and maintained consistent reporting SLA adherence across high-volume daily batch processing cycles.

• Supported end-to-end ETL workflows in Hive and Big Query for reporting and analytics teams. • Maintained incremental and historical datasets using Apache Hudi, ensuring accuracy and preventing data duplication datasets. • Troubleshot Spark jobs and Kafka stream failures to reduce downtime for critical analytics pipelines.