Sep 2025 – Present · 9 mos · Hyderabad

• Building production-grade GenAI systems and data infrastructure at enterprise scale.
• GenAI & Multi-Agent Systems
• Built Sprint Insights, a multi-agent LLM platform (LangGraph) that automates sprint reporting for 10+ agile teams - eliminating 3–4 hours of manual work per cycle with 90% report accuracy.
• Developed a Contact Center Analytics agent enabling natural language querying over structured data, replacing SQL dependency for business users. Includes RAG pipeline, LLM-driven SQL generation with validation loops, and hallucination monitoring.
• Data & ETL Platform Engineering
• Architected a cloud-agnostic ETL-as-a-Service platform (Python, Apache Beam) supporting on-prem, Kubernetes, and GCP deployments — scaled to 100+ production pipelines processing terabytes daily with a 60–70% reduction in time-to-production.
• Led re-architecture of a legacy 10+ year-old SSIS reporting system, cutting critical report generation from 8–9 hours to under 2 hours (75% speedup) while ensuring regulatory SLA compliance.
• Reliability & Infrastructure
• Built a priority-based resource orchestration service to prevent CPU/memory exhaustion across 100+ concurrent ETL jobs on fixed-capacity Windows infrastructure.
• Developed a self-healing virtual environment wrapper that reduced environment-related pipeline failures by 96%, now adopted as the enterprise deployment standard.

Aug 2023 – Sep 2025 · 2 yrs 1 mo · Hyderabad

• Cloud-Agnostic ETL-as-a-Service Platform - Flagship Enterprise Initiative
• As Technical Lead on one of the company's top 5 strategic initiatives, I partnered with senior architects and principal engineers to design and build a revolutionary configuration-driven data integration platform. This became the central nervous system for all enterprise data movement.
• What makes it special:
• > Platform-agnostic architecture: Built with Python and Apache Beam, deployable seamlessly across on-premise Windows servers, Kubernetes clusters, and Google Cloud Platform - eliminating vendor lock-in and enabling cloud migration flexibility
• > Zero-code pipeline creation: Business analysts can now deploy production-ready data pipelines in hours by simply configuring metadata tables - no development required
• > Unified batch & streaming: Single framework handles file systems, databases (SQL Server, BigQuery), event streams (Kafka, Pub/Sub), REST APIs, with built-in export/delivery capabilities
• > Enterprise-grade reliability: Comprehensive auditing, alerting, data quality validation, intelligent retry mechanisms, and exactly-once delivery semantics
• Results:
• > 60-70% reduction in time-to-market for new data pipelines
• > Powers 100+ production pipelines processing terabytes of data daily
• > Simplified pipeline creation across the Contact Centre Space
• > Significantly reduced infrastructure costs through unified codebase
• > Eliminated dedicated ETL developer dependency for standard use cases

Aug 2022 – Aug 2023 · 1 yr · Hyderabad

• Legacy ETL Modernization Led the complete re-architecture of a decade-old SSIS-based ETL framework that generates time-sensitive regulatory reports for banking operations. Working with a system I was completely new to, I designed and implemented a modern master-slave architecture from the ground up.
• Impact:
• > Reduced report generation time by 75% (from 8-9 hours to under 2 hours)
• > Prevented compliance violations and financial penalties through timely report delivery
• > Eliminated critical operational gap by introducing on-demand data reload capability a long-awaited feature that replaced manual date-hardcoding processes
• > Implemented business-defined data quality rules with automatic quarantine, maintaining process continuity despite bad data
• > Built real-time monitoring dashboards providing single-pane visibility for platform support and business users The framework now supports heterogeneous data sources with simplified integration patterns and seamless Tableau connectivity.