// how I think
Engineering
Philosophy, field notes, and technical writing. An engineer's lab notebook — built in public.
// core principles
Engineering Philosophy
Systems Thinking
Problems should be solved at the system level. Local optimization fails. You fix the symptom and break something upstream. A misaligned pump bearing isn't a bearing problem — it's a vibration problem, which is a mounting problem, which is a structural problem. The engine room taught me to trace every failure back to its root, not its expression.
Operator-Built Software
The best maintenance software I've ever seen was written by someone who had to use it at 3am in a storm. Software built by operators, for operators, solves different problems than software built by engineers who've never run the system. My competitive advantage is that I've been on both sides of that divide — and I build from the operator side.
Overbuilt by Design
Maritime engineering taught me that reliability isn't a feature — it's a design philosophy. Redundancy, safety margins, and robust failure modes aren't overhead. They are the system. A ship's main engine doesn't get a patch Tuesday. When you build for environments where failure is catastrophic and a technician isn't always available, overbuilding becomes the only rational choice.
Applied AI vs. Theoretical AI
I don't care about AI in the abstract. I care about AI that can tell a maintenance technician which bearing is about to fail, and why. Operational automation. Decision assistance. Knowledge that doesn't walk off the ship when someone retires. The question I ask of every AI system: does it make an operator's job better in the field, right now, under real conditions?
// field notes
Engineering Log
What Ship Engineering Teaches You About System Reliability
Redundancy, safety margins, and why the engine room is the best classroom for distributed systems thinking. If a network outage costs you money, imagine a main engine failure in the middle of the Pacific.
Building an AI That Knows Your Equipment
Why knowledge ingestion is the hardest part of industrial AI — and how DigitalRound approaches the problem of encoding 30 years of an engineer's expertise into a system that doesn't retire.
Why Operator-Built Software Wins
The gap between software built by operators and software built for operators is wider than most engineers realize. Notes on how the best industrial tools get made — and why user research isn't enough.
The Maintenance Knowledge Problem
Every industrial facility has the same problem: knowledge lives in people, not systems. What happens when that person leaves, retires, or just isn't available at 2am when the generator alarm trips?
Redundancy as a Design Principle
On maritime engineering's approach to failure: design so that any single failure doesn't take down the system. Applied to software architecture, startup operations, and everyday decision-making.
From Engine Room to AI Agent: Same Problems, Different Stack
The engineering problems I face running an AI agent fleet are structurally identical to running a ship's engine room: resource management, failure isolation, watchkeeping, and keeping the whole system alive under load.
// reference material
Technical Notes
Maritime
Chief Engineer Operations
Watch schedules, maintenance planning, ISM compliance, and engine room management on ocean-going vessels.
AI Systems
Agent Architecture
Multi-agent orchestration, context management, model selection, and operational AI deployment patterns.
Hardware
Embedded & IoT
Raspberry Pi, Arduino, industrial sensors, MQTT, edge compute, and physical-digital system integration.