← Back to projects
Internal PlatformSelf-serviceInfrastructure as Code

Kafka → Terraform migration

Company
Toptal — Fashion Digital client
Role
Senior Site Reliability & Platform Engineer
Period at the company
Aug 2025 — Present

Problem

Kafka topics, ACLs and connectors were managed through an aging internal deployment tool with no version history, plan diff or safe review path before production. This created operational toil and concentrated deployment knowledge in a small group.

Approach

Working with SRE and Cloud Engineering, I co-architected the import of existing resources into Terraform state without downtime. Targeted imports and drift detection reconciled code with the running environments; pull-request reviews and automated Terraform plans then became the self-service release path for future changes.

Impact

  • Kafka resources across environments became version-controlled, reviewable and reproducible.
  • The outdated deployment application was retired in favour of an automated infrastructure pipeline.
  • Teams can promote changes from a readable Terraform plan instead of relying on manual, undocumented operations.

Tech stack

TerraformKafkaGitHub Actions