Articles, short research notes, and talks.
The long-form argument behind every project here: memory substrates with provenance, capability-scoped authorization, and observability indexed by decision rather than span.
I built hooks that stop a coding agent from opening unreviewed pull requests or merging unchecked code. Review rounds got past them with ordinary shell. What failed, the rule that held, and the fix of mine that reopened the hole.
An indexer looks simple until the chain it indexed a minute ago changes. How I would structure one for an EVM chain, who owns progress, what happens during a reorg, and the protocol facts behind each decision.
AI agents raised productivity and moved the attack surface with it. Where it moved, which controls reduce the loss, how to see in near real time when an agent goes wrong, and what building these controls taught me.
I tested a design where agents hold task-scoped signing keys, while a separate executor holds service credentials and checks each action. A live model run faced a planted prompt injection and revealed an unexpected environment leak.
How I designed and rolled out a GitOps delivery platform for 35+ services and 10+ teams, what the trade-offs were, how teams came to trust it, and what broke along the way.
Why retrieval-augmented generation is not enough, and what a real memory substrate for autonomous agents might look like.
Metrics, logs, and traces were built for human operators. Autonomous agents demand a new observability primitive: the reasoning trace.
Static allow-lists will not keep autonomous agents safe. Governance must be dynamic, uncertainty-aware, and grounded in simulated impact.
Coordination, conflict resolution, and shared context are the defining challenges of multi-agent systems.
What a decade of production infrastructure engineering taught me about building agentic systems.
MCP is a promising step toward standardized tool use, but it leaves hard questions about state, identity, and trust unanswered.
Stuffing history into a prompt is a hack, not an architecture.
Benchmarks that measure single-turn correctness miss the point. We need to evaluate agents across episodes, not just turns.
A systems-level view of the infrastructure needed for trustworthy autonomous agents.