May 2025 – Nov 2025 · 6 mos · Gurugram, Haryana, India · Hybrid

• Individually developed the Overwatch Anomaly Detection system, a product for real-time detection of transaction anomalies using TimescaleDB, Redis, Kafka, and OpenSearch.
• Collaborated with the Platform team to build internal tools and monitoring systems using Grafana and OpenSearch Dashboard for real-time observability and anomaly detection.
• Contributed to the development of dashboards and alert systems using Grafana to monitor system anomalies, latency,and error patterns.
• Developed automation scripts and backend utilities using Python, optimizing workflows such as log analysis and exception detection.
• Implemented Drain3 AI with Elasticsearch for automatic log template mining, enabling detection of unseen and new error patterns in real-time.
• Designed and tuned TimescaleDB hypertables with continuous aggregation for minute/day-level transaction metrics,boosting query performance by 10x for historical success-rate analysis.
• Automated dynamic thresholding pipelines using rolling 15-day baselines, enabling adaptive anomaly detection without manual intervention.
• Developed alert prioritization logic based on impact (volume, duration, deviation), reducing noise and minimizing false-positive alerts.
• Integrated multi-channel alert routing (Telegram, Email, Teams) with severity-based escalation policies, ensuring timely stakeholder notification.