Prannay Khushalani

Engineered a greenfield retail data platform from 0 to 1, developing high-volume ingestion pipelines using Scala and Spark to process millions of raw user behavioral records into structured, analytics-ready tables. Architected "pre-machine learning" data foundations by building robust ETL workflows that deliver clean, feature-ready datasets for retail predictive modeling and customer signaling teams. Optimized large-scale batch processing jobs in the Hadoop ecosystem, implementing broadcast joins and custom partitioning to reduce processing latency for multi-terabyte retail transaction datasets.

Data Engineer - Albertsons Companies

(2023-12 - 2026-03)

• Engineered a greenfield retail data platform from 0 to 1, developing high-volume ingestion pipelines using Scala and Spark to process millions of raw user behavioral records into structured, analytics-ready tables.
• Architected "pre-machine learning" data foundations by building robust ETL workflows that deliver clean, feature-ready datasets for retail predictive modeling and customer signaling teams.
• Optimized large-scale batch processing jobs in the Hadoop ecosystem, implementing broadcast joins and custom partitioning to reduce processing latency for multi-terabyte retail transaction datasets.
• Developed standardized, granular financial datasets for downstream reporting and AOP planning, utilizing SQL and PySpark to unify data across disparate retail verticals and ensure metric consistency.
• Deployed a production-grade LLM-powered monitoring system using LangChain, automating the triage of pipeline failures and reducing mean time to resolution (MTTR) by 70%.
• Integrated automated data quality gates using Great Expectations, enforcing schema conformance and validation checks at every stage of the retail data lifecycle to ensure high-fidelity reporting.

Master of Science - Business Analytics and Artificial Intelligence - The University of Texas at Dallas

Bachelor of Technology - Information Technology - Vishwakarma Institute of Technology