AUTOMATED AWS COST OPTIMIZATION SYSTEM (EC2 SCHEDULING + EBS SNAPSHOT LIFECYCLE MANAGEMENT)
Designed and implemented a serverless cost optimization system managing 15+ EC2 instances and 50+ EBS snapshots using AWS Lambda and EventBridge. • Automated EC2 start/stop scheduling based on instance tag policies, which helped to reduce unused compute costs by 30–40% across non-production environments. • Implemented EBS snapshot lifecycle management to identify and delete unused snapshots, which reduced storage costs by 25%, • Enforced IAM least-privilege access and integrated CloudWatch logging/alerts, achieving secure and observable operations with zero manual intervention
: LOG ANOMALY DETECTION & ALERTING SYSTEM (DEVOPS / AIOPS)
• Built a containerized log analysis service that processes 10,000+ logs/day based on past time-window data. • Detected anomalies with 85%+ accuracy using statistical thresholds and Isolation Forest model. • Reduced false-positive alerts by 40% using alert suppression and stateful tracking. • Deployed on Kubernetes with scalable ingestion, handling 3x traffic spikes without affecting performance.
AI ASSISTED DOCKER MONITORING SYSTEM
Built an NLP interface using the Docker Engine API to translate plain-text queries into structured commands for fetching real-time logs and container states. • Implemented health checks based on CPU, memory, and restart patterns, improving issue detection time by 50%. • Designed workflow