Recruiter-question FAQ

Are you a programmer or a manager?

Both, in the way that matters for the role I want. I write production code, I ship infrastructure, and I run a team that ships in manufacturing. I am not the right hire for a deep-systems ML research role; I am the right hire for a senior operator role where being able to read, write, and ship code is what makes the management actually effective.

What is the difference between Allmine and Tamko corporate?

Allmine Paving is a subsidiary inside Tamko Building Products. I joined Allmine as Quality and Raw Materials Manager in 2022, was promoted to Operations Manager at Allmine, and then promoted into the corporate AI and Six Sigma Lead role serving Tamko’s full footprint of nine plants. Tamko is a privately held industrial business in the multi-billion-dollar roofing manufacturing space.

Why are you looking?

The Tamko corporate role is going well. The playbook is shipping across plants. The next stretch for me is a senior operator role at a company where AI is part of the core operating value, not a corporate side-program. Either a manufacturer or process operator that is serious about its operational AI layer, or an AI company that needs someone to run the operational fabric around the product.

Why should an AI company hire you over an ML engineer?

You should not, if you are hiring an ML engineer. Hire me for the role where someone needs to ship the model into a real environment, build the operational discipline around it, run the rollout, handle the customer or operator experience, and make sure the system is still working six months later when nobody is watching. That role is not the same role as building the model.

Why should an industrial company hire you over a traditional ops PM?

Because I can actually read and write the code that gets deployed. The traditional ops PM hands the model to engineering and the deployment to IT and hopes everything composes. I have run AI rollouts where I was the person debugging the controller, tuning the model, scripting the alarm logic, and standing on the floor with the operator at 0300 when the system did something unexpected. Most industrial PM hires cannot do that.

What is your security clearance situation?

I held the appropriate clearances during my Army career. I do not currently hold an active clearance. I am US-eligible, US-citizen, and would not have a clearance reciprocity issue if a role required one.

Do you have a degree?

Two from Embry-Riddle Aeronautical University, both completed during active military service. MS Unmanned Systems and BS Aeronautics.

Where do you live and will you relocate?

Harpers Ferry, West Virginia. Eastern time. I am not actively trying to relocate. The family is settled, two young kids, and the timing on a move is not great. I am happy to do remote-first with travel or a hybrid pattern to an East-coast hub.

What kind of company size fits?

Five to thirty people on my direct team. The company itself can be much larger; I just want a team I can shape end-to-end without three layers of permission between me and the work.

What is your day actually like?

Mornings start with the personal Overseer system briefing me on what shipped overnight, what is blocked, and what needs my approval. Then I am usually in a mix of plant Teams calls, model and rollout work, code review, and the kind of operator-facing problem-solving that happens when a deployed system surprises you. Afternoons skew toward writing — proposals, playbooks, the documentation that makes a rollout durable. Evenings are family, and then maybe an hour of the personal-agent system work that keeps me current.

What is the Overseer system on your work page?

A personal multi-agent operating layer I built and run for myself. Ten production services across a Mac and a VPS. Postgres-backed event bus. A remote MCP server reachable from Claude Desktop and ChatGPT custom connectors. About 17,500 durable memories. Sixty-five-ish proposed actions a week, all gated by one-tap approval. It is the operating posture I bring to industrial AI, at personal scale and live every day.

How do you handle a model that drifts?

Drift detection runs as a service from the day a model is deployed, not as a quarterly review item. When it triggers, the on-shift engineer gets paged and the system falls back to the operator-known-good control profile until a human signs off on the new behavior. The trust is in the watcher, not the model.

How do you handle an AI rollout that the plant operators do not want?

Slowly and on their terms. The first system has to make the operator’s day measurably easier without changing their authority. Once they trust the first one, the next ones get easier. Rolling out AI against operator skepticism by mandate is how every failed deployment I have ever seen started.

What stack are you most comfortable in?

Python for ops and core logic. TypeScript for edges. Postgres as the default state store. Cloudflare for edge. Astro for content surfaces, Next.js for operator consoles. Claude for reasoning, OpenAI for embeddings and voice. MCP as the integration surface for anything that needs to be reachable from Claude Desktop or ChatGPT.

What do you not want?

A role whose first ninety days are “AI strategy” with no plant, customer, or product attached. A pure-consulting engagement. A team where the model is the marketing and operations are an afterthought. Any role where the success metric is a tweet count or an investor slide.

What hobbies do you have?

Family is the main one. Wife and two young kids. Outside that: aviation-adjacent stuff still lives rent-free in my head, occasional outdoor work, and the personal-agent system itself — which is half hobby and half operating leverage.

How fast can you start?

Fast, with a respectful notice period to Tamko. Two to four weeks once an offer is in hand.