At PhonePe, I build data platforms that move roughly 3 TB of change events daily from 150+ MariaDB shards to downstream ML and analytics teams. I engineered self-healing for 200+ Spark Streaming jobs, optimized a critical PySpark job by 40%, and built validation workflows that catch data-quality issues before they reach users.
Previously at Bare International, I designed real-time Kafka, Redis, and MariaDB processing for over 1M events per four-hour window, built a Delta Lake data lake that reduced query latency by about 45%, and created a graph-based data catalog for 40+ datasets.
I also bring hands-on experience building KPI reporting and automated data-quality checks, along with practical AI infrastructure work including a RAG triage chatbot and an open-source governed AI data analyst.
