I help companies take back control of their data and AI. I design and build private cloud platforms, on-premise AI inference stacks, and automated security pipelines so organizations can run sensitive workloads, fine-tune models on proprietary data, and meet compliance requirements without depending on third-party vendors.
Running databases on Kubernetes has a reputation for being painful. I built CloudDB to change that: a Kubernetes-native, self-hosted DBaaS control plane with a React dashboard and a Go API that wraps the Zalando Postgres Operator. Here's what I built, why, and where it's heading.
Most Machine Learning projects never make it to production. Not because the models don't work, but because building a model is only 10% of the journey. The other 90%? That's MLOps. This article explores how MLOps bridges data science and engineering, covering DevOps foundations, MLflow, and Kubeflow.
eBPF evolved from the original Berkeley Packet Filter, which primarily handled packet filtering. Now, it lets you run custom programs in the Linux kernel without changing its source, hooking into events like network packets or system calls. This unlocks powerful capabilities for monitoring, filtering, and even modifying system operations.