AgentBuild - Production AI for Enterprise Practitioners

Most AI projects don’t fail in the lab. They fail in production.

AgentBuild is a weekly newsletter for data and AI practitioners - architects, engineers, and technical leaders - who are past the demo stage and dealing with what actually comes next. Evaluation gaps. Decision traces. Agent infrastructure that breaks at scale. The gap between what your vendors promise and what your systems deliver.

If you’re still figuring out what an LLM is, this isn’t for you. If you’re the one in the room responsible for making AI work in the real world, it is.

What you get

Every issue covers one problem from production AI in depth - frameworks, patterns, and hard-won lessons from shipping AI systems inside large organisations. No hype. No roundups. No “top 10 tools” lists.

Topics include evaluation-first AI design, decision trace architecture, agentic system choreography, data readiness for AI, Enterprise AI tranfomration frameworks.

About Sandi

Global CTO and enterprise AI/data leader. I write about taking agentic AI from prototype to dependable, governed production systems, particularly in complex and regulated organisations. I’ve spent 18+ years inside Databricks, AWS, the NHS, and UK FSIs - watching enterprise transformation on data & AI projects succeed and fail in the same predictable ways.

AgentBuild is where I think out loud about the patterns that matter.

Join the crew

If you're building production AI systems and want frameworks that survive contact with the real world - subscribe.

Be part of a community of people who share your interests. Participate in the comments section. Connect, because that matters in learning.