Davi Kawasaki
Building cloud platforms, GitOps delivery systems and resilient infrastructure that help engineering teams ship safely.
About
I'm a Senior Platform Engineer and Site Reliability Engineer with hands-on experience across Kubernetes, Terraform, AWS, GCP, GitOps, CI/CD, observability and Infrastructure as Code. I've worked with startups and multinational organisations across LATAM, the US and EMEA.
Recent work includes reducing Kubernetes observability costs and resource consumption by 20%, cutting disaster recovery from hours to under 10 minutes, reducing Artifact Registry costs by 19%, resolving more than 50 container CVEs and removing 40% of stale Kafka infrastructure.
Read the full resume →Tools & Skills
Technologies and practices I use to build reliable platforms and reduce developer toil
Infrastructure & IaC
Platform, IDP & GitOps
Observability
Software, Data & Security
Selected engineering case studies
Internal platforms, reliability automation and developer tooling with measurable outcomes
FisiHome — take-home physiotherapy SaaS
Kawasaki Web Soluções
Multi-tenant rehabilitation SaaS that connects prescribed video exercises with patient recordings and asynchronous professional review.
Read case study →ZyfinBot — Global personal finance SaaS
Kawasaki Web Soluções
Multi-tenant personal-finance SaaS that captures expenses through the official WhatsApp API and supports secure adviser access across currencies.
Read case study →Kafka → Terraform migration
Toptal — Fashion Digital client
Migration of legacy Kafka resources into version-controlled Terraform pipelines that removed approximately 40% of stale infrastructure and retired the old release tooling.
Read case study →GitOps deployments with multi-image Kustomize
Toptal — Fashion Digital client
Python/Flask Internal Developer Platform refactor that added secure multi-image deployments, cut CD time by about one hour per day and reduced UI load time to about 350 ms.
Read case study →Latest posts
Practical findings on infrastructure, platform engineering and developer experience
How to delete cloud artifact registry container images after one year excluding live ones
A registry policy based on age and position can delete an image that production still pins. This cleanup job reads the deployment state first, stops when its input is unreliable, and publishes its plan days before it deletes anything.
Architecture Decision Record: promote microservices with pull requests to a GitOps repository
A sanitized Architecture Decision Record from an internal developer platform: every promotion is written as a pull request, and a controller reconciles the cluster from the repository.
Need a more reliable platform?
Let's discuss self-service infrastructure, automation and the operating practices that help engineering teams move safely.