Wiktor Kisielewski

• Worked with the team to bring separate monitoring systems into one platform for metrics, logs and traces, handling terabytes of telemetry daily.
• Helped define observability standards for 800+ microservices in the cloud and on-premises, including bare metal. Introduced wide events and shared metadata conventions with the team, then worked with other teams to put them into practice.
• Responsible for incident management and on-call for trading services, including SLOs for critical operations, blameless postmortems and runbook automation to shorten recovery time.
• Added JVM metrics and traces to Java services using OpenTelemetry auto-instrumentation, without application code changes.
• Added monitoring for PostgreSQL and Elasticsearch, including slow queries and replication health, plus Kafka consumer lag.
• Built self-service alerting so developers and QA can manage their own alerts and thresholds. Added reports showing each team's alerts, incidents, metric coverage and log usage.
• Built a RAG tool for searching postmortems and internal docs during incidents, a scanner for sensitive data in logs, and an LLM tool that flags incidents in team chat before they are formally reported.
• Plan disaster recovery and run failover drills for trading infrastructure.

• Sole DevOps engineer for mortgage systems with bank integrations across multiple countries. Ran the AWS and GCP infrastructure, including Kubernetes, GitOps deployments, Terraform, Ansible and monitoring.
• Led the migration from AWS to GCP, reducing infrastructure costs by ~40%.
• Designed the disaster recovery process and tested it by restoring the full production environment several times.
• Handled the infrastructure work for ISO 27001 certification.
• Switched releases to Git tags and set up deployments so QA could deploy to test environments themselves.
• Ran chaos experiments to check how the platform handled infrastructure failures.

• Automated customer installations on their own infrastructure with ArgoCD and Terraform, bringing setup time from days to minutes.
• Automated recovery from recurring failures, reducing out-of-hours alerts.
• Ran Kubernetes in the cloud and on-premises (vanilla, OKD, Rancher, k3s) for an AI/NLP platform, including GPU nodes, model serving and data pipelines.

• Built Azure pipelines that let other teams run routine tasks without DevOps help.
• Automated Azure DNS and Private Link synchronization between regions with Azure Pipelines and PowerShell.

• Brought monitoring into one setup and standardized bare-metal Kubernetes provisioning with Ansible.
• Maintained a multi-TB Elasticsearch cluster and Java services, with 24/7 on-call coverage for customer cloud and data center deployments.

Monitored infrastructure and troubleshot hardware in WP's data centers.