PrakharMittal

IDFC FIRST Bank

Data Solution Analyst

• Leads end-to-end development of scalable ETL pipelines on AWS EMR, PySpark and Jupyter, processing high-volume banking datasets with ~30% better processing efficiency.
• Spearheaded the migration of enterprise workloads from AWS S3/EMR to on-prem HDFS, cutting cloud infrastructure cost by ~25% while tightening governance.
• Orchestrates dependencies through Apache Airflow DAGs, reducing manual intervention by ~70% and improving scheduling reliability.
• Optimized PySpark transformations — partitioning, caching, join tuning — for a ~35% cut in job runtime and better cluster utilization.
• Built interactive validation frameworks in Jupyter that reduced data-validation effort by ~40%.
• Implemented reconciliation logic, data quality checks and monitoring that lifted data accuracy and consistency by ~85%.
• Partners with business stakeholders on scalable data solutions, accelerating delivery timelines ~20% for critical reporting.
• Holds ~99.9% SLA adherence across production data pipelines through proactive monitoring and issue resolution.

ABSA · Barclays Africa Group

Senior Data Engineer — on-site with LTIMindtree

• Initiated and led the Data Rail Program, moving legacy Teradata warehouses onto a Hadoop architecture and translating SQL to Spark SQL — 45% faster query processing.
• Managed 12 Big Data engineers across on-site and offshore teams at >98% delivery compliance.
• Converted Teradata marts and extracts for 12 African countries into Hadoop marts with Hive, Spark and Teradata utilities — 50% faster ingestion, 30% savings on infrastructure and licensing.
• Built multi-source business and financial analytics pipelines that doubled ABSA's data-driven revenue potential through better decision-making.

LTIMindtree

Specialist — Data Engineering

• Supported 100+ enterprise-scale ETL applications across the African continent on Hadoop, Teradata, Informatica, TWS Tivoli, CA WADE and Airflow at 99.9% SLA.
• Led the Informatica 10.1 → 10.4 Virtual Server migration, tuning mappings, sessions and pushdown queries for a 40% gain in processing speed on high-volume banking data.
• Designed ETL pipelines with data quality rules, partition pruning and join optimizations, improving accuracy 80–85% across reporting and compliance systems.
• Deployed Python, Spark, Sqoop and Shell scripts integrating structured and semi-structured data into Hadoop for near-real-time fraud detection and customer insight.
• Refactored legacy models with partitioning, bucketing, clustered tables and a Spark→Tez engine switch — 35% faster queries, 25% better cluster utilization.
• Drove the move of on-prem Hadoop data marts to Dell ECS object storage via HDFS–S3A DistCp, tiering, lifecycle policies and JCEKS-encrypted credentials — 45% less on-prem storage overhead, eleven-nines durability with geo-replication.
• Onboarded 10+ business units with ingestion frameworks and Ranger/Atlas governance, lifting enterprise Hadoop/ECS adoption 50%.