Mikael Lirbank is an independent software engineer and AI systems specialist who helps teams build and ship robust, reliable, high-quality AI products. With more than a decade of experience in predictive analytics, machine learning, and agentic systems, he works with venture-backed startups, venture capital firms, and enterprise clients. His work spans AI architecture and development, developer tools, production readiness, evaluation systems, and data workflows.

Projects include coding-agent plugins and skills for Convex and Neon, an email-based supplier agent for Coupa, data anonymization workflows for Databricks Lakebase, and AI products for procurement, health benefits, and communications. Earlier, he created Conservator, a data lifecycle platform for neural-network training that became an OEM product for Teledyne FLIR. Lirbank also runs STΛR MODΞ, a collective of independent builders developing open-source projects, developer tools, and commercial products.

His own work includes Neon Testing, a utility for integration testing with Neon Postgres. He occasionally speaks at meetups, conferences, and corporate events about using AI to build better software