Jagadeesh Kumar Email & Phone Number

Greater Chicago Area

Riverwoods, Il, Us

- In the role of Senior Principal Software Engineer at Discover, I lead the engineering for Emerging Technologies group's Centralized Logging Service(CLS) and responsible for end to end technical ownership, strategy and roadmap, performance improvement and optimization.- I lead the design and implementation of processes in CLS that are responsible for streaming/processing/persisting logs across the entire organization(Payments/Card/Banking/Analytics and other business units). CLS handles daily 6+ billions of logs, 18 Terabytes of data, with end to end latency of 1 sec(including data transfer) using tools Kafka, Nifi, Logstash, Elasticsearch, Kibana, AWS , Machine Learning, Terraform, Ansible, Java, Junit, Gradle, Python, Ruby, Docker, Kubernetes, OCP, Painless - Complete ownership of In-house cost effective custom solutions (which replace AWS managed services like Lambda, Kinesis Fireshose, Kinesis) that meets aggressive performance, throughput and latency requirements- Compute Optimization : With performance improvements in the application and runtime that I made, reduced the application infrastructure hosted on AWS from 283 m5.xlarge to 123 m5.xlarge ec2's.- Machine Learning to identify/notify the changes in the volume trends and respond with actions

Chicago, Illinois, Us

Design, implementation & testing of Infrastructure-layer’s services that form the basis for business modules in the hybrid trading system's built using Java, native C programming , shell , perl etc.Evaluate the interfaces between hardware and software and adapt the software to the hardware or to upgrade the interfaces to improve the performance.Analyze system capabilities to improve the response time latency. Implement the Enhancements & Bug fixes for core infra components and tech tools that are part of very large scale distributed computing environment which requires high performance, ultra-low latency, and high availability, 24*5.5. Different roles in engineering/management.Metrics#Failover : 3 seconds(Slow) vs 150 millis(Fast) for the trading engine failover.New in house Transport: Very Few millis(CORBA) vs 100 micros vs 60 micros(w/ solarflare)New in house pub/sub messaging: Few millis(ActiveMQ) vs 150-250 micros(including transport). This messaging system is Kafka like and has improved features(dynamic filtering) which traditional Kafka doesnt support.

Zurich, Ch

Design, implementation of backend server side distributed processes that collaborate for the calculation of margin and assess counterparty's aggregate exposure and collateral needs across business lines in global markets.Performance improvement, architectural changes, middle ware use/upgrade etc.

Boston, Ma, Us

Hyderabad, Telangana, In