dm3ch

- Implemented company-wide SSO, MDM and an access portal, enhancing security and reducing operational overhead. - Migrated from git-encrypted secrets to HashiCorp Vault + Vault Secrets Operator, unifying secret management and improving rotation workflows across clouds/clusters. - Achieved ~40% AWS compute cost savings using Karpenter with Spot instances, Istio AZ-affinity, resource tuning and autoscaling. - Refactored the Istio service mesh to simplify management and remove custom components. - Administered Kubernetes clusters using Karpenter, EKS, Talos and FluxCD. - Deployed centralized logging with Grafana Loki. - Worked with developers on performance bottlenecks and troubleshooting. - Performed security vulnerability analysis and remediation. - Implemented alerting via Grafana IRM and VictoriaMetrics. - Introduced granular infra access with SSO and Gravitational Teleport. - Contributed to web3 RPC proxy development. - Led resource usage analysis and optimization. - Migrated applications and CI builds from x86_64 to ARM64. - Created reusable GitHub Actions workflows, shared actions, standardized Dockerfiles and automated stale branch/PR cleanup. - Coordinated with external vendors (web3 RPC, cloud, SaaS) for integrations and service quality. - Performed DevOps Tech Lead responsibilities for 1.5 years, improving infra reliability and CI/CD workflows. - Provided interim Team Lead coverage during vacations and a 3-month sabbatical. - Gained experience in hiring, interviewing, onboarding and offboarding.

• Redesigned and participated in rebuilding of CD processes - decreased amount of tools used during deployment from 10 to 5, unified CD processes for apps, reduced amount of k8s IaaC sources. This allowed our engineers to operate infrastructure more efficiently and greatly reduce onboarding time for new engineers. • Started project of integration of modern observability tooling (VictoriaMetrics, Loki, Tempo) - for example migration from ELK to Loki allowed us to store 6-12 months of logs for the same price as 1 month before. • Reworked existing staging environment IaaC and tooling - decreased new env creation time from a day to 30 minutes. • Reworked some DBs on stages to k8s operator-controlled solutions to improve stability. • Involved in simplification of CI processes - decreased amount of abstraction layers in CI, making CI code more user-friendly.

• Built HA Gitlab company-wide instance in k8s - improved availability from 85% to 99%, migrated gitlab to k8s, decreased upgrade process time from 1 month to 15 min. • Implemented CD process for application in k8s, later redesigned it to allow faster and more stable deploys in different points of presence - allowed parallel deployment, reduced deploy time from 30m to 30s for one app. This allowed us to move from a single deploy per day to more smaller one. • Participated in expert group on building company-wide multi-tenant monitoring solution with longterm storage - allowed us to store 1 month history for all our clusters. Reduced storage price from 20k (Prometheus per cluster) to less than 3k (single VictoriaMetrics cluster). • Designed and participated in the implementation of a company-wide docker registry. • Developed custom kubernetes scheduler for internal purposes - make dynamic stagings/autotests 15-20% less cost (cluster size about 1.5k of CPU cores), due to more efficient scheduling (make one env/test-run use as less unique nodes as possible, that makes node termination faster). • Participated in the design and implementation of a dynamically scaled selenium grid (up to 1k CPU) in GKE, which allowed us to improve browser test stability and lower the amount of time needed to run the test pack. • Take a part in open source tools development, and contributed patches upstream for such projects as nginx-ingress, gitlab, etc - this prevented the creation of big amount of self-developed tools.

• Migrated project from Openshift to Kubernetes - allowed us to adopt existing k8s operators and charts (in 2018 most of projects wasn’t ready to work in OpenShift). Make us able to easier find new colleagues and faster onboard them. • Migrated project from own custom manifest generation tooling to helm and helm file. • Completely rebuild CD process - allowed to reduce delivery time from 12h (half-manual approach) to 5m (fully automatic). • Integrated Grafana Observability Stack (Cortex/Loki/Tempo) and Sentry error tracking solution - greatly improved debugging process by cross linking metrics/logs/traces from same UI. • Designed and involved in implementation of secret management flow - almost all CI/CD processes have used fully dynamic auth for accessing secrets in Vault.

• Researched service meshes and FaaS tools. • Built internal tools project with Gitlab AutoDevops and GCP to allow developers in the company to deploy self-built tools without any OPS help in typical cases. Integrated this platform with company security auditing tools - allowed us to fully automate 90% of new tool hosting requests and greatly decrease time for MVPs of internal tooling. • Prepared PoC and participated in the migration of frontend builds from TeamCity to Gitlab CI. It allowed us to significantly reduce build queue time from about 1h to 1m by using autoscaling gitlab runners in GKE.

- Maintained client servers and the applications deployed on them. - Deployed clients' applications, including working with clients on their apps architecture. - Configured and built infrastructure using Kubernetes and virtual machines. - Gained hands-on experience with a wide range of software tools (e.g., ClickHouse, Elasticsearch, Redis, RabbitMQ) due to the company’s outsourcing focus.