Human Still
I’m Tushar Murudkar. I write about what it takes to build and run reliable systems as more of the work gets automated.
A failed deployment can generate plenty of signals. The harder work is deciding what happened, what to do next, who should make the call, and how to tell whether the fix worked. That’s where AI, systems thinking, and engineering leadership meet for me.
Human Still is for engineers, platform teams, and engineering leaders making those decisions. I share field notes from software delivery at scale, experiments with AI in operations, and ways for teams to learn from failures and decide where human judgment matters most.
Start here
Autonomous DevOps — How to decide what an AI system can act on, where it needs approval, and how to verify the result.
What 23,000 Deployments Revealed About Software Delivery at Scale — What deployment data can tell us about how a delivery system actually works.
Turning Deployment Failures into Operational Learning — How an AI bot helps a delivery platform learn from failed deployments.
Subscribe to get new field notes in your inbox.

