Developed a production-ready, self-hosted LLM deployment stack for the opf/helm-charts repository, enabling scalable Kubernetes deployments with Helm and APISIX as the AI gateway. The solution introduced a configuration seeder job to automate APISIX route and consumer setup, streamlining onboarding and ensuring reproducibility. Integrated observability using Prometheus and Grafana, providing dashboards and metrics to monitor LLM throughput and latency for development environments. Enhanced deployment flexibility supported both local development and external AI provider routing. Contributed to chart refinement, documentation, and examples, demonstrating strong collaboration and maintainability. Core technologies included Kubernetes, Helm, APISIX, Prometheus, Grafana, and Python.