SHERVIN GHANBARY

Joined the DevOps Engineering team.

• Ansible and AAP: Developed and deployed an Ansible script via Ansible Tower to automate the creation of Vault AppRoles, directories for secrets, and associated infrastructure. Integrated Ansible variables with an inventory list from ServiceNow, successfully migrating existing jobs from Tidal automation tool.
• Terraform and Vault: Created and implemented Terraform scripts to build HashiCorp Vault infrastructure, including policies, secret engines, access methods, and AppRoles. Restructured Vault architecture using Vault namespaces, simplifying secret management processes and improving scalability and security.
• Cloud & On-Prem Support: Provided ongoing support for Elastic Cluster issues, ensuring high availability and resolving performance problems. Assisted in AWS account provisioning, creating new IAM roles, IAM policies, and managing VPC networking to optimize cloud resource management.
• Python Scripting: Designed and implemented a script to dynamically identify and update all repositories containing Cortex YAML files, pushing automated commits to fix issues with the group provider configuration, streamlining CI/CD processes.

Sole member of Cloud Engineering team in London, responsible for the AWS infrastructure.

• Terraform & CloudFormation: Developed modules and templates for AWS infrastructure, deploying many AWS services such as VPC, EFS, Lambda, APIGW, ECS, ALB, ect. I've also used Packer to produce custom AMI's.
• CI/CD with GitLab: Developed many python scripts used for automation, including automating AWS tasks (e.g., RDS Instance migrations to Aurora cluster, ECR image pushes, unit test creation, terraform deployments through python api endpoint for etc.) via GitLab CI pipelines.
• Support Duties: Provided production cloud support during UK hours, resolving AWS deployment and infrastructure issues. Utilized CloudTrail to look for changes to resources, CloudWatch for troubleshooting issues in ECS tasks, errors in containers, api-gateway execution failures... etc.

Key Projects:

• AWS Infrastructure to Azure Migration: Migrated infrastructure from AWS to Azure for Azure clients. Moving from ECS to AKS.
  Developed Terraform modules for Azure resources (Virtual Networks, Hadoop, Blob Storage, Load Balancers, DNS).
  Deployed containers on AKS using Helm, including pods, ingress controller, ingresses. Private link to access dashboard through private IP's.
  Built python scripts to sync market data from AWS S3 to Azure Blob Storage and to migrate ECR images to Azure Container Registry.
  
• Disaster Recovery (DR) Plan Implementation: Developed regional DR solutions for RDS through python scripts run by GitLab jobs to restore snapshot, and automated backups to secondary region.
  Automated infrastructure provisioning and recovery processes for disaster recovery. Created configuration in all envs to support deployments to another region creating. blue/green deployments. Created support in our python api for blue/green deployment.

• Lead Production Upgrade Pipeline Optimization: Led a team of 3 to reduce production upgrade time. Reduced pipeline duration by 30% from 2 hours. Created python scripts to split Docker image deployments to AWS ECR via GitLab CI across many runners using gitlab "matrix". Improved Terraform deployment speed by separating static resources like ECS clusters from dynamic infrastructure, like tasks and services..

• Single Sign-On (SSO) via Cognito & Okta: Developed automated Okta user creation via Lambda & python in CI pipeline.
  Managed user roles, groups, and passwords through python-based Lambda functions.
  Created ALB and Route53 records for Okta UI access; automated environment migration to Okta using python and CloudFormation.

• Fleet all of our EMR deployments, and prepare AMI, for EMR 7.x.x: Configured all CloudFormation templates, for emr fleets to add resiliency to deployments. Updated all prod environment configurations to pick up fleet changes. Created bootstrap actions in bash script to be run using packer to produce custom AMI's for EMR 7, allowing SSH for Hadoop user. Created python integrations tests to run via gitlab CI to test future changes to EMR. Made EMR bootstrap scripts backward compatible with EMR 6.x.x to not affect prod environments.

Worked in the Platform Engineering team, focusing on automation, infrastructure as code (IaC), and CI/CD pipelines for various projects.

Key Projects:

• Jira Integration with Lambda, SQS, and API Gateway: Provisioned Created AWS Lambda, SQS, and API Gateway through Terraform within deployment through Drone-CI pipeline to enable bi-directional communication between Jira instances across different teams.
  Developed API endpoints to trigger Lambda function that replicate incident tickets between Jira's, ensuring secure, policy-compliant ticket management.
  
• Kubernetes Deployment Pipeline for AWS EKS: Built CI/CD pipeline in Drone to automate deployment on AWS EKS.
  Created Kubernetes yaml files to deploy, services, secrets, and cron jobs, as well as custom container monitoring pods to view all pod stats and post to slack. Responsible also for carrying out rolling updates to our managed pods.
  
• GitLab Sync Pipeline for On-Prem to AWS Migration: Created CI/CD pipeline to sync GitLab repositories between on-premise infrastructure and AWS, supporting the decommissioning of on-prem servers and migration to GitLab on AWS.
  

Support Duties:

• Provided L1/L2 platform support, resolving high-priority incidents on our on-prem servers, and our EKS clusters.
• Debugged issues in CI/CD pipelines, container deployments, and EKS environments.

Other Contributions:

• Automated workflows and created monitoring scripts using Bash, PowerShell, and Python for platform management, migrations of data across to other servers, and reporting container stats to Nagios.
• Developed custom dashboards and health/status monitoring for microservices, on-prem servers, and Docker containers, with alerts sent via Slack and Nagios, created python script to monitor servers and to post status to confluence via confluence api.