Austin, Texas, United States
Master's in Computer Science from California State University -Los Angeles Skill Sets:- Programming Languages/Provisioning :- Python, Shell Script, C++ , Ansible Benchmark :- FIO, IOmeter, Iperf, SGEMM, GPU stream, Resnet50, LLM fine tunning Tools/Framework :- PowerBI,KVM,QEMU,Wireshark,Docker,Jira Monitoring Tools :- Grafana, Prometheus, Loki, Alertmanager, Webhookd, Collectd, Intel XPUM, Promtail Profiling Tools :- Vtune, Intel XPUM
1) Delivered performance optimization for large language models and recommendation systems on AMD GPUs. 2) Improved MLPerf Inference results through architecture-aware tuning, benchmarking, and performance analysis. 3) Lead the development of automated CI/CD infrastructure for continuous benchmarking and performance validation.
1)Developed and maintained Kubernetes manifests and Helm charts with ArgoCD, automating the deployment and scaling of microservices for ML/AI applications. 2)Architected and deployed telemetry solutions for Intel’s AI GPU clusters using Kubernetes, enabling efficient capture and analysis of telemetry metrics and logs. 3)Leveraged DeepSpeed for distributed inference, enabling the deployment of large models across multi-GPU setups to achieve fast inference times and reduced operational cost. 4)Automate container testing for performance and validate the cluster at scale.
1)Extending the CPU/GPU optimization work scope at scale to the AI domain. 2) Using docker compose architected cluster monitor architecture with service components of frontend (Grafana, Prometheus, Alertmanager, Webhookd and Loki ) and backend (Collectd, Promtail, Intel XPUM) services. 3)Developed scripts and an exporter using the Redfish API to collect and monitor BMC information for enhanced system management and oversight. 4)Implement proactive alerting systems to quickly respond to critical telemetry events using Prometheus Alertmanager and Webhookd. 5)Implemented and Developed Dockerfiles to optimize AI cluster docker images using Github Actions
1) Monitoring Jenkins builds, build failure debugging and work with the peer teams to resolve the issue. 2) Ran AI/ML (Dgemm, RocHPL, HPL, RESNET50) workloads on AMD Mi200-250 GPUs.
1) Development of test automation methodologies for Xeon server CPUs. 2 )Data Visualization - Ownership of PowerBI tool to create reports and dashboards. 3) Daily monitoring and debugging of 500+ prototype XEON servers. Including troubleshooted various complex issues related to test execution.
1)Installation, Configuration and Upgrade of Linux 5/6/7, SUE-11/12/15, CentOS. 2)Full automation with no manual interference using Jenkins and Python in order to run CI testing framework. 3)Scripting and automation using Python and Shell Scripts for testing and benchmarking purposes. 4)Responsible for configuration and setup of virtualized OpenStack setup using KVM and QEMU on Intel Server Platform. 5)Experience with OpenStack in deploying large pools of compute, storage and network resources via CLI and Horizon Dashboard 6)Running and Debugging Performance Benchmark to measure Packet Loss in the cloud Environment 7)Responsible for configuration and setup of Wireguard in the Kernel on Intel’s platforms. 8)Performance monitoring, benchmarking Wireguard in Kernel and in Docker container on Intel’s gen by gen CPUs. 9)JIRA usage for bug tracking and maintaining records/tickets
1) Manged shelving of book in chronological order. 2) Maintained library's database by barcoding books 3)Provided monitoring and analytical support for the database