Leading the design and development of Layer 7 security and networking solutions for Edgio CDN.Built and managed Kubernetes clusters on on-premises infrastructure to host high-performance applications including ClickHouse, Elasticsearch, Spark, and Airflow, optimizing resources for data-intensive and machine learning workloads.Led the design and development of Edgio’s native bot management solution and API security solution powered by machine learning. Implemented advanced JA4+ networking fingerprinting on the CDN and designed scalable pipelines to deploy ML model outputs across more than 15000 servers globally.Implemented seamless deployment of WAF ruleset updates across the entire edge infrastructure, introducing a monitoring, alerting, patching, and testing framework to ensure uninterrupted services and prevent accidental outages.Built a fully remote team of data scientists and ML engineers to build next-gen security solutions using machine learning.

Built the security team from 4 engineers to 14 engineers dedicated to developing next-generation layer 7 security solutions for Edgio CDN. This includes overseeing the creation of Bot Manager, WAF, API security, stateful rate limiting, and leveraging machine learning algorithms to improve these products

Built a data ingestion pipeline in Golang that parsed CDN logs in protocol buffer format, handling 100 million+ logs per hour and inserting them into an Elasticsearch cluster.Developed an open-source library, WAFLZ, for WAF and rate limiting, and integrated it with Nginx and other webservers.Worked towards making the deployment of services cloud-native and integrating them with edge computing.

Re-designed GCS backend for drone deliveries, achieving 50X performance improvement by migrating to C++ and Azure.

Contributed to open-source WAF engine in C++, implementing zero-copy optimizations to scale performance by 20X.Utilized GDB, flame graphs, gperftools, and ASAN for bug fixes and performance improvements.Designed rate limiting solution to prevent DDOS attacks, improving counting accuracy by 50% on a large-scale distributed system of 20,000 serversImplemented a versioning system for customer configurations and an IAM framework. Built REST API server in Golang with AWS RDS as backend db.

Developed custom web server modules in C for Nginx and Lighttpd webserver to efficiently throttle bit transfer rate over a specified port and send custom logs over the wire using protocol buffers.Designed and developed security and access layer for various services on file server using LUA scripting language embedded in Nginx webserver. Wrote applications using Openssl to secure keys and certificates on the server.Developed an analytics tool for gathering passenger usage statistics on an aircraft using C, C++, BOOST, socket programming, TCP/IP and shell scripting. Cross-compiled this tool for various architecture viz.: x86, ARM, ARM-Freescale, MIPS, and PPC.

Dependable computing systems research laboratory, University of North Texas Created and managed a cloud with four servers and multiple client instances on XEN hypervisor 3.1.2 and host OS withCentOS 5.4 over Linux kernel 2.6. Implemented algorithms for better resource management. Current research is in powermanagement and anomaly detection in cloud computing . Used Hadoop MapReduce as a test bed for all the experiments.Technologies: XEN hypervisor, CentOS, Hadoop, Map reduce.