Outstaffing to UK/US product companies via ventionteams.com; promoted from Junior to Middle DevOps Engineer.

Fanatics Collectibles — Fanatics’ sports collectibles business, spanning Topps, Fanatics Live, and Fanatics Collect and serving millions of collectors worldwide. (fanaticscollect.com, fanatics.live). ● Improved release reliability and reduced manual deployment effort by modernizing 11 application pipelines and migrating critical workloads to Amazon EKS, as part of 85+ improvements across AWS infrastructure and delivery operations. ● Brought manually managed AWS resources across five accounts under a centralized Terraform/Terragrunt repository—including EKS, networking, DNS, databases, ECR, IAM, and RabbitMQ configuration—and developed and maintained internal Terraform modules for project-specific use cases, making provisioning repeatable and environments more consistent. ● Enabled private connectivity from AWS organization accounts to GCP load balancers by integrating Megaport connectivity with AWS PrivateLink VPC endpoints. ● Reduced load-testing environment preparation time from 30–45 minutes to 2 minutes by building an automated, containerized framework integrated with CircleCI and EKS, lowering costs by preventing idle resource usage and simplifying scalable testing across multiple applications with Helm-based controls. ● Improved release reliability by implementing Argo Rollouts with active and preview environments, automating load and integration tests before promotion, and enabling rollback to reduce the risk of customer-facing downtime. ● Managed EKS node provisioning and disruption policies for diverse application workloads using Karpenter NodePools and EC2NodeClasses deployed through Argo CD; piloted amd64-to-ARM infrastructure migrations, reducing development amd64 node count by approximately 20%, with broader ARM adoption projected to lower node compute costs by 50%. ● Reduced Datadog APM trace noise by filtering unnecessary traces, cutting indexed spans by approximately 1 million per month. ● Automated off-peak scaling for Aurora Serverless PostgreSQL using Kubernetes and Terraform, reducing minimum capacity by 50% for six hours nightly across Live and Collect services, with projected annual compute savings of $21K. ● Validated an RDS MySQL 8.0-to-8.4 upgrade strategy for databases with manually configured replicas across AWS accounts, using Bash/Python monitoring scripts to identify replication failures during Blue/Green testing and verify replication recovery during in-place upgrades. ● Reduced the risk of production application downtime by migrating OpenSearch from a single-node setup to 3 data nodes across three Availability Zones and optimizing shard and replica configuration. ● Enabled teams to validate image delivery before production releases by deploying automated image resizing, format conversion, and caching in staging using AWS Lambda, S3, and CloudFront. ● Strengthened security across 6+ application services and supporting infrastructure by centralizing secrets in 1Password, hardening container configurations for 8 services, and securing credential management in CI/CD pipelines.

Crown Business Finance — UK-based finance brokerage connecting businesses with lenders for loans, asset and vehicle finance, refinancing, and tax funding. (crownbusinessfinance.co.uk). ● Maintained uninterrupted service during migration from WordPress/Wix to Amazon EC2 by developing the migration plan and provisioning the required AWS infrastructure. ● Resolved routing conflicts between WordPress and the frontend after migration by configuring NGINX to direct requests to the correct application. ● Designed environment-specific configurations (e.g., dev, staging, production) using GitHub Actions and AWS services, enabling safe testing and seamless transitions between deployment stages.

Renta (renta-cars) — Uzbekistan-based digital car-rental platform that lets users find, book, and manage vehicle rentals online. Contract & Freelance Engagements (remote). ● Enabled faster delivery of updates by reducing CI/CD pipeline execution time from 10 minutes to 3 minutes—a 70% reduction—through build, release, and deployment optimization. ● Reduced API response times from 10 seconds to 1–2 seconds by implementing Mobile caching for frequently requested data, reducing backend load and improving stability during traffic spikes.

UIC Group — Uzbekistan-based software development and IT outsourcing company that has delivered 200+ digital products to the local market, including web, mobile, and business-management solutions. ● Standardized deployments for 100+ projects by migrating from manual, script-based processes to containerized environments using Kubernetes, Helm, Docker, Docker Swarm, GitLab CI/CD, and Harbor. ● Reduced deployment workflow template complexity from 20+ stages to 4 while supporting 74,000+ pipeline executions per month. ● Restored reliable access to external APIs by resolving production IP-blocking issues through a Docker-based proxy network, ensuring uninterrupted service despite regional restrictions. ● Reduced data-loss risk and manual backup effort by automating daily PostgreSQL backups across multiple VMs to a dedicated backup server using Python, Bash, and Ansible, enabling recovery on a replacement instance if the original VM failed. ● Implemented observability using Prometheus, Grafana, Node Exporter, and Uptime Kuma across 150+ microservices and products, reducing MTTR from 60 minutes to under 10 minutes. ● Shortened CI/CD execution time for frontend projects by 40% through dependency caching and build optimization, enabling faster releases. ● Avoided an estimated $10K in annual hosted-runner costs by deploying self-hosted GitLab Runners across our on-premises infrastructure. DOMO (domo.uz) — Uzbekistan’s household utility-payment app, bringing gas, electricity, water, internet, and waste-service payments together in one platform with currently 2.8M+ downloads nationwide. ● Supported platform growth to nearly 1 million users by scaling application infrastructure with Kubernetes and Docker Swarm. ● Maintained reliable service during peak demand, supporting 5,000+ concurrent requests with stable performance. ● Reduced database query latency by 40% through PgBouncer tuning and connection-pooling optimization, helping the application respond faster. ● Increased database read capacity and reduced primary-node load by automating read-traffic distribution across two or more PostgreSQL replicas, depending on demand. Xolodilnik (Internal project) — Online grocery platform for fresh-food delivery, with personalized product discovery, order tracking, and integrated tools for order and delivery management. ● Migrated Odoo ERP from a single-instance deployment to a scalable Kubernetes environment, enabling the application to run across multiple replicas to handle increased demand. ● Established capacity-based scaling for Odoo by analyzing resource consumption and request-handling capacity across 6 workers per pod, using measured performance data to configure Kubernetes Horizontal Pod Autoscaler and plan capacity for peak demand. ● Created reusable Helm and GitLab CI/CD infrastructure templates to optimize releases and pipeline performance across repositories for future Odoo projects. ● Designed a 14-week DevOps mentoring program at UIC group; mentored one DevOps intern to production readiness.