Skip to main content

Creating Your Own Role

5 min read

Eng Manager

New roles emerge from need. If someone proposes 'AI Product Engineer' and can back it up, consider it.

Tech Lead

You see the gap. Propose the role. Show the impact. Leadership listens to clear business cases.

Tpm

New roles need a scope. You can help define: what does this person own? What's the success metric?

Creating Your Own Role

TL;DR

  • Sometimes no job title fits. You're doing X, the company needs Y, and the org chart has a hole.
  • Proposing a new role = making a business case. Problem, solution, impact, why you.
  • It works when: there's real need, you've already been doing it, and leadership trusts you.

"AI Product Engineer" and similar roles didn't exist in 2022. People proposed them. Companies created them. You can do the same.

When It Makes Sense

  • You're already doing the work. You've been building AI features, no title. Propose the title.
  • There's a gap. "We need someone to own AI in product." Nobody owns it. You propose yourself.
  • External hires are expensive or slow. Creating an internal role is faster. You're already there.

The Proposal

1. Define the problem. "We're adding AI to the product ad-hoc. No ownership. Inconsistent quality and cost."

2. Define the role. "AI Product Engineer — owns AI feature development from spec to ship. Bridges product, engineering, and AI."

3. Define success. "Ship 2 AI features per quarter. Reduce LLM cost by 20%. Define our prompt management process."

4. Define why you. "I've built X and Y. I've been doing 60% of this already. I want to own it formally."

5. Define the ask. "Create this role. Title me appropriately. Let me hire/contract for support if needed."

Making It Stick

  • Document everything. What you're doing. Impact. Hours. Show the gap between "current state" and "need."
  • Get a sponsor. Your manager, a director, or a PM who sees the need. They advocate.
  • Start small. "Let me do this for 3 months. If it works, make it official." Proof before commitment.
  • Use external validation. "Companies are hiring for this. Here are 5 job posts." You're not inventing from scratch.

When It Doesn't Work

  • Company is in freeze. No new roles. Wait or leave.
  • Political resistance. "That's engineering's job" or "That's product's job." Find the ally who sees the gap.
  • You can't prove impact. Do the work first. Then propose. Reverse order rarely works.

You're building AI features. No title. Ad-hoc. Inconsistent quality. Nobody owns it. You're doing 60% of the work already.

Click "Proposed role" to see the difference →

Quick Check

When does proposing a new role make sense?

Do This Next

  1. Map what you're already doing — List tasks that don't fit your current title. Is there a pattern?
  2. Find the business case — What problem does this solve? What's the cost of not having it?
  3. Draft a one-pager — Problem, role, success, you. Share with your manager. "Can we explore this?"