I’m fully immersed in all things AI at the moment, especially LLMs and the infrastructure and tooling that help integrate them into every conceivable workflow, as well as the rapidly evolving field of AI Engineering more broadly. One of the things I am currently spending time on is how LLM-enabled AI systems can reason over real-world infrastructure and produce answers people can trust. At Kinara Systems, that work takes shape in sauble.ai, where we are exploring how agentic systems can help IT and network operations teams connect operational data, investigate incidents, and explain likely causes with supporting evidence.
I’m equally interested in how AI is changing software engineering and technical product development itself: how teams design, build, test, and verify systems as AI tools reshape product development. That includes coding agents, agentic workflows, tool integration, evaluation, and observability, as well as the practical question of how people learn to work effectively with these systems.
This notebook is a record of some of that work, along with whatever else catches my attention.