Vinay Surapureddy

Applied research role building production-grade MLOps and cloud infrastructure for AI systems on AWS spanning Kubernetes provisioning, CI/CD automation, RAG/LLM pipeline deployment, and Explainable AI (XAI) inference layers. - Designed an Explainable AI (XAI) inference layer over a Reinforcement Learning (Q-routing) SDN system on AWS EKS, using SHAP-based surrogate modeling to make AI routing decisions transparent and enterprise-auditable. - Deployed a RAG (Retrieval-Augmented Generation) pipeline with vector embeddings and LLM inference endpoints for AI-driven knowledge retrieval; tracked all experiments with MLflow model registry for versioning and reproducibility. - Built GitHub Actions CI/CD pipelines for staged model promotion (dev→staging→prod); provisioned Kubernetes with Terraform and Helm (RBAC, namespace isolation); deployed Prometheus + Grafana with SLI/SLO-aligned drift alerting. - Automated cloud cost governance with Python (boto3), reducing idle resource spend by 25% across multi-region AWS environments.

Led cloud infrastructure, CI/CD ownership, and DevSecOps for Waste Management's enterprise AWS environment which is one of North America's largest environmental services companies, supporting millions of customers across multi-region production workloads. - Architected multi-region AWS infrastructure (VPC, EC2, S3, IAM, EKS, RDS) with Terraform and CloudFormation IaC — achieving 99.9% uptime SLAs. - Designed Jenkins and GitHub Actions CI/CD pipelines with Groovy Shared Libraries across 15+ microservices with improved release frequency by 30%, reduced post-release defects by 25%. - Deployed containerized workloads on Amazon EKS with Helm and Docker; implemented blue/green and zero-downtime deployments with Karpenter auto-scaling and RBAC policies. - Authored Ansible playbooks and Bash/Unix scripts for OS patching and configuration management across 100+ Linux servers with reduced manual effort by 40%. - Integrated SonarQube SAST + Trivy scanning (shift-left DevSecOps), aligned with SOC 2 and CIS benchmarks with reduced vulnerabilities by 45%. Deployed Prometheus + Grafana observability stack, reducing MTTR by 35%.

Infrastructure and DevOps support for Kaiser Permanente - one of the US's largest nonprofit healthcare providers — covering AWS/on-prem provisioning, Azure DevOps CI/CD pipelines, Python/Shell automation, Linux server administration, and production incident response. - Built Azure DevOps CI/CD pipelines for Azure (App Service, AKS) with automated test gates, canary/blue-green strategies, and rollback triggers. - Developed Python and Shell/Unix automation reducing operational toil by 35%; managed Jira sprint boards for Agile release coordination. - Implemented Prometheus, Grafana, and CloudWatch monitoring with SLI/SLO dashboards enforcing high availability SLAs. - Managed Linux/UNIX administration (RHEL/CentOS): ELK Stack log analysis, security hardening, patch management, and disaster recovery runbook maintenance. - Led Sev1/Sev2 incident response, RCA, and blameless post-mortems with reduced recurring incidents by 30% over 6 months.

Graduated with Master’s degree from NAU