Help organizations learn a new way to work.

What this is not

This is not primarily an AI engineering role. It is not prompt engineering, conventional implementation consulting, or process automation with a new name. An Arc Practitioner works at the intersection of people, work, systems, knowledge, and intelligence.

What you'll do

The interesting problems are rarely cleanly defined. You'll learn how work actually happens, then help shape an Arc around it.

1. Discover and decompose organizational workflows.
2. Identify where AI can carry specialty work and where human judgment should remain.
3. Design the context, tools, knowledge, permissions, and boundaries intelligence needs to work effectively.
4. Help turn transient conversations and outputs into durable organizational knowledge.
5. Establish appropriate approval, governance, and evidence around consequential actions.
6. Work with technical teams to connect existing systems rather than unnecessarily replacing them.
7. Test whether the resulting environment actually produces the intended outcome.
8. Make the whole thing understandable to the humans who will live with it.
9. Help those humans develop their own practice around it.

The implementation matters. So does what happens six months later.

What we're looking for

One strong discipline. Enough curiosity to cross into the others.

• You can follow information through a messy organization and figure out where reality diverges from the diagram.
• You're technically fluent enough that APIs, permissions, data structures, Git, command lines, containers, and AI models aren't mysterious, even if you aren't the person who builds every component yourself.
• You can sit with an executive, an engineer, a security team, and the person who actually performs the work, and understand that each of them knows something the others don't.
• You document decisions and expose assumptions.
• You know the difference between an impressive demonstration and an operationally sound system.
• You're comfortable saying that something shouldn't be automated.
• You can use AI seriously as a working partner without surrendering judgment to it.
• When you own an outcome, you understand that ownership includes the consequence.

AI expertise

You do not need to be an AI expert. You need practical fluency with contemporary AI systems and a willingness to develop much more of it. We'll care more about whether you can reason about where intelligence belongs, what context it needs, what authority it should have, how its work should be evaluated, and what evidence should survive, than whether you know every current framework.

Models will change. The practice has to survive them.

How we work

At amotivv, people hold outcomes rather than task lists. AI already does meaningful work inside the company: it researches, recalls, drafts, analyzes, critiques, and carries specialty load. Humans provide judgment, relationships, direction, and accountability for the consequence.

You won't be joining a conventional consulting organization with a finished methodology. You'll help develop the practice. We expect Arc Practitioners to bring what they learn in the field back into the discipline: patterns that worked, assumptions that failed, new methods, better explanations, and evidence that changes our own thinking.

What good looks like

Not "adopted AI." Useful work happening differently.

• People have greater capacity without losing responsibility.
• Intelligence can do meaningful work without being given ambiguous authority.
• Organizational knowledge compounds instead of disappearing into conversations and individuals.
• Important actions leave evidence.
• The environment remains understandable.
• The organization becomes increasingly capable of operating it without depending on us.

Experience

No single required background. Show us what you've actually been responsible for.

Scopes of practice

Responsibility, not tenure. These aren't tenure bands; they're scopes of responsibility. Someone new to Arc may already bring decades of judgment from another field.