Pranav Tavva

I’m a Data Engineer focused on building scalable, cost-efficient data pipelines on AWS. I enjoy turning complex data workflows into reliable, production-grade systems that improve performance, reduce spend, and stay resilient under load. I take ownership end-to-end—from pipeline redesign and migrations to monitoring and failure handling.

At Zenoti, I re-architected data pipelines from Iceberg/PySpark to a MySQL-backed Anchor + Athena approach, cutting processing time from 23 hours to 7 minutes (99.4% faster) and reducing cost from 100 DPU to 0.4 DPU (99.7% reduction). I automated Gross Margin (GM) reporting using AWS Batch + .NET, eliminating 10–14 hours of manual effort while improving financial data accuracy. I also optimized large-scale fact table processing by shifting heavy deletes/upserts to Athena, reducing runtime from 2+ hours to 30–45 minutes and compute from 50+ DPU to 1 DPU (5.4K yearly savings).

I’ve led critical ingestion and reliability work, including AWS → Azure migration with dual ingestion and zero downtime for consistent reporting. I built an API-based ingestion system to push Data Lake data into core systems, improving fault tolerance and scalability, and strengthened pipeline stability with monitoring, retry mechanisms, and failure handling. On the project side, I built an Athena-powered Data Lake query platform (with Redis-backed authentication) and a real-time Kafka + Spark Streaming pipeline processing 50K+ events/min with watermarking and checkpointing for fault tolerance.