Now

Currently focused on

• Writing about running AI in production — infrastructure, retrieval, and the decisions behind shipping agents.
• Building and running several production AI agents on my own self-hosted infrastructure.
• Running a personal AI assistant lab for code review, ops, and content drafts.

Recently shipped

• Moved my AI stack from a single server to a three-tier setup — dev, production, and a shared AI + data layer.
• Site rebuild around writing rather than agent demos.

Currently learning

• MCP server design patterns at scale — discovery, versioning, security boundaries.
• Hybrid retrieval (vector + lexical + rerank) trade-offs in mixed-language corpora.
• Cost and performance tuning of self-hosted inference and embeddings.

This page updates whenever my focus shifts — usually monthly. Inspired by the now page movement.