Ivan Malianov

Data Platform Reliability Engineer

Operate data middleware for a billing transformation serving 15M+ subscribers, with more than 500M incoming Kafka messages per day.

• Stabilized 24 Java CDC deployments through JVM/GC tuning, OpenShift resource sizing and dataflow isolation; observed 2-3-week runs without recurring configuration-related failures or unplanned restarts between scheduled deployments.
• Implemented PostgreSQL HA connectivity with Patroni, etcd and HAProxy, corrected ScyllaDB seed and replication-factor configuration, and validated failover and recovery scenarios.
• Built observability for all existing dataflows across 18 OpenShift services and 4 Jenkins job groups for HDFS, using Kafka-derived delivery, throughput and backlog metrics, a production Rust exporter and Bash/curl connector checks.
• Estimated detection time improved from 1-6 hours via BI symptoms to 15-30 minutes, and component localization from several hours to 20-40 minutes. Created 15+ dashboards and migrated 25+, covering 500+ metric names/recording rules and 400+ alerts.
• Restored connectors and data flows and contributed to resolving approximately 300 defects through RCA and fix validation. Diagnosed Elasticsearch failures associated with an OpenShift CSI driver, applied a workaround and escalated to Red Hat support.
• Automated configuration corrections, storage cleanup, operational checks and deployment log collection with Python and Bash; supported release workflows in Jenkins and GitLab CI/CD.
• Rotated identified plaintext secrets and moved them to Jenkins credentials and Vault/OpenBao; migrated service-account SSH access to keys and helped remove shared/obsolete accounts.

Selected AIOps projects at Beeline

• Data Quality Test-to-Incident Agent: built a Python MVP matching 400+ tests with Jira history. Recommendations led to 40+ incidents created manually by specialists and approximately 100 linked tickets. Estimated manual matching took 1-2 hours; reviewing agent-generated results now takes 10-15 minutes.
• Log Pre-Analysis Agent: designed the architecture and guided two developers on a read-only Python/LangGraph MVP with a custom MCP server. It searches Elasticsearch and Jenkins logs and suggests causes and next checks; piloted with four administrators on one production system.

DevOps Engineer | Feb 2023 - Sep 2024

• Migrated 20+ critical servers, including Kafka, PostgreSQL, GitLab and Vault, from on-premises infrastructure to Yandex Cloud using Hystax, with no data loss and several hours of database downtime.
• Built GitLab CI/CD for Docker/Compose applications with automated tests, quality gates and security scans; automated deployments, backups and key management.
• Containerized Pentaho Community Edition to simplify upgrades; implemented Prometheus/Grafana monitoring for earlier issue detection and faster incident response.

Python Developer | Nov 2021 - Jan 2023

• Built Python tooling for 200+ specialists and 400+ Hadoop clusters, integrating Ambari/Ranger APIs. Added Celery task-existence checks, connect/read timeouts and retries with exponential backoff and jitter.
• Helped bring the service from MVP to production within one year. Automation reduced manually processed L2 requests by 40% versus pre-implementation levels, according to ITSM statistics; cluster health checks enabled early intervention on incident risks.
• Mentored two interns who completed the program and received offers to join the team.