Kirk Abbott

Cloud security as a discipline is mature. The cloud-and-AI intersection is not. Most of the cloud security tooling that exists today was designed for human-driven workloads — shared responsibility models, IAM patterns, network controls, all assuming humans are the ones writing the requests. AI agents change that assumption. The least-privilege patterns we built for humans don't compose well with agents that need broad action surfaces but produce unpredictable inputs. The audit trails that work for human-paced activity don't scale to agent-paced activity. And the threat model expands: prompt injection, model exfiltration, and adversarial inputs become security concerns nobody had to think about three years ago.

That's the gap I want to spend the next decade in. Cloud architecture and security on one side; AI/ML systems on the other. They don't combine cleanly today, and the people who can think rigorously about the seam between them are going to be in short supply for a long time.

Most of my engineering experience came from building web applications and learning the cloud as I went — the kind of work where you eventually realize the security model matters more than any framework choice. Pages, APIs, queries, deployments; you ship enough of those and you start asking “what happens when this is wrong?”

I'm currently in the U.S. Air Force doing cyber defense operations, which puts me on the other side of the same problem — defending systems other people built, often without the context the builders had. Both views point at the same gap: builders ship systems they can't fully reason about; defenders inherit systems they can't fully see.

I also run a small SaaS on my own infrastructure — multi-service stack, real users, real payments, real OAuth integrations — which keeps me close to the operational realities of the things I write about here.

Cloud architecture, security, and AI/ML are where I want to spend the next decade. The work I'm publishing here is part of that bet — research, projects, and writing that compound over time.