✓ Built 5+ RAGs for Gen AI based application security review to reduce manual investigation time by 40%, from weeks to days

✓ Developed knowledge base pipeline to collect data and create Vector DB on Opensearch, increasing scan accuracy by 30%

✓ Simplified data retrieval for RAGs working across 3 teams to reduce execution cost by 94% (weekly 84 Glue Jobs to 5 Spark Jobs), reduce storage cost by 99.99% (TBs to MBs) and operational cost by 15% by reducing 10+ AWS Infra Components

✓ Implemented a fine grained access control framework for UI authorization using OpenPolicy agent, scalable up to 10,000 TPS

✓ Leading the architecture for Amazon's Threat Detection Platform, processing 5PB datastreams per day in real-time

✓ Led the design and migration of a monolithic application into 5+ distributed and scalable microservices using AWS

✓ Implemented auto-scaled system to increase throughput from 10 TPS to 700 TPS and reducing operational cost by 81.3%

✓ Developed a self-service platform to achieve 99% reduction in data onboarding time, with ingestion latency of 1 GB/min

✓ Designed a datastore with DynamoDB Index pointing to files stored in S3 buckets, reducing the storage cost by 90.8%

✓ Implemented an aggregation engine using Flink to handle 3,000 events/sec, reducing time to detect from 2 hours to 2 mins

✓ Automated the kernel patching process for an EMR Flink cluster across 200+ EC2 instances by transferring the distributed state to S3 prior to patching, ensuring 0% data loss and reducing manual effort by 96.86% (from 8 hours to 15 minutes)

✓ Re-designed S3 file reader to support 100GB/min, reducing CPU usage by 70% resulting in 60 % reduction in EMR infra costs

✓ Implemented a multi-threaded routing service to handle 30,000 events/sec with average latency of 150ms using SQS - ECS

✓ Developed an in-house YAML domain-specific language, enabling security engineers to author 1300+ threat detections

✓ Re-architected and developed the threat detection system, reducing end to end latency by 93% (from 15 mins to 1 min)

CS Grader for CS 4301 Cluster Computing