The Missing Layer Between AI Agents and Company Data
I am building Mainmind: a governed knowledge layer that gives people and AI agents the same account of how a company works, without trapping that knowledge inside any model or application.
Writing
Essays on agent experience, AI systems, and leadership.
Long-form notes for builders and leaders turning emerging capability into reliable software: agent experience, evaluation, enterprise platforms, distributed systems, and the judgment behind technical decisions.
- AI Agents
- AI Engineering
- Agent Experience
- Architecture
- DevOps
- Distributed Systems
- Engineering Judgment
- Enterprise AI
- Evals
- Harness Engineering
- Knowledge Architecture
- Leadership
- LoopOps
- Mainmind
- Operating Model
- Organizational Design
- Performance
- Product Engineering
- Product Judgment
- Reliability
- Runtime Design
- Telemetry
- WebAssembly
Fussy Systems Need Fuzzy Harnesses
We keep bolting rigid systems around agents. The better architecture is a firm organizational core with a disposable, evolving harness around it.
Every Application Has Two Surfaces—and Agent Evals Should Diagnose the Second
Agent experience is not an alternative to agent evals. It is the product surface those evals should help us diagnose and improve.
DevOps Is Moving Toward LoopOps
DevOps made shipping and operating software a continuous discipline. LoopOps is the next step: making improvement itself a first-class, measurable loop for humans and agents.
Telemetry Should Close the Improvement Loop
Telemetry should not end at observability. As software teams add coding agents and agentic workflows, telemetry needs to become the evidence layer that helps products get better.
Agent Experience Is Bigger Than Developer Experience
More tokens, more context, and more connectors do not automatically create better AI products. Agent experience borrows from user experience and developer experience, but agents need a working environment for every kind of operational task.
What Senior AI Engineers Actually Own
The senior part of AI engineering is not knowing every model release. It is owning the product, system, risk, and organizational judgment that turns model capability into durable software.
A Practical Operating Model for Enterprise AI Agents
Enterprise agents need more than tools and prompts. They need ownership, evaluation, permissions, observability, and a rollout model that matches the risk of the workflow.
The Hidden Cost of Abstractions: When Elegant Code Becomes a Performance Trap
Every abstraction is a trade-off. Here's how to identify when your beautiful, clean code is silently destroying performance—and what to do about it.
Why I Fell in Love with WebAssembly (and You Should Too)
WebAssembly isn't just another compile target—it's a paradigm shift that's quietly revolutionizing how we think about performance, portability, and the future of computing.
Mental Models for Distributed Systems: Thinking Beyond Single Machines
Distributed systems break our intuitions about computing. Here are the mental models that help me reason about complex, distributed architectures without losing my sanity.