yaku-cloud-platform

Architecting YAKU: Engineering Governance-as-Code

At Bosch Software Flow, I served as a Senior Software Engineer for YAKU, a platform designed to replace manual spreadsheets with automated, verifiable compliance. I was responsible for the core API architecture and the cloud orchestration layer that allows YAKU to execute "Autopilots"—specialized micro-services that collect and evaluate evidence from various DevOps tools.

The Core Orchestration

The heart of YAKU is its Core Engine, which I architect in Go. This engine handles the lifecycle of a "Run," managing the execution of multiple Autopilots and consolidating their results into a standardized evidence format.

API First: I designed the Core Interfaces to ensure a clean separation between the orchestrator and the execution environment.
Autopilot Ecosystem: I developed the framework for Autopilots, enabling them to run as ephemeral workloads that interact with the YAKU API to report compliance status.
Distributed Execution As we required a workflow execution engine we did not rebuilt one ourself but stuck with Argo Workflows.

Cloud-Native Delivery via GitOps

To support the scale required by Bosch, I engineered the deployment strategy using a modern Cloud-Native stack:

Flux & Helm & Kubernetes: I supported designing and implementing the GitOps workflow, ensuring that the YAKU infrastructure and its various microservices are deployed declaratively and remain in sync with the repository.
Scalable Execution: By leveraging Kubernetes, the platform can dynamically scale the execution of Autopilots, providing immediate feedback on quality gates during the CI/CD process.

Impact & Retrospective

YAKU represents a shift from "checking boxes" to "checking code." By building a robust API and a reliable orchestration layer, we enabled teams to automate their release evidence. This project allowed me to combine my experience in high-stability systems (from my automotive background) with the agility of modern cloud-native platform engineering.