The roleAI only compounds when it's woven into how an organization actually works, not bolted on as another tool. We need a systems thinker who will work alongside teams across the company (sales, success, ops, finance, and engineering), see how work actually gets done, and remove the friction. The toolkit is broad and the right answer depends on the problem.
You'll own the AI adoption layer that sits across the whole org, with engineering as a first-class user. Making engineers measurably faster with AI tooling is one of the highest-leverage applications of this role, and it's where you'll often start. The mandate is broad: make AI the default way work happens here, with whatever combination of tools, agents, and systems that takes.
What good looks like- Our internal AI assistant is the first stop, not the last resort. People use it because it works. Information flows are mapped, gaps get followed up on, and answer quality improves week over week.
- Every team has at least one custom agent that earns its keep. You've sat with the team, watched them work, and shipped something narrow that takes real toil off their plate.
- Engineers feel measurably faster. Agent-augmented engineering (Claude Code, Cursor, agent code review, agent QA) is the default workflow, not an experiment. PR cycle time and time-to-first-deploy for new hires move in the right direction.
- Documentation is alive. It updates when the code changes. When it's wrong, an agent flags it. When somebody asks a question the docs should have answered, the docs get better.
- Adoption is visible. We can see who's using what, where the holdouts are, and what's actually moving the metric. The "I'm bought in" claim gets backed by data.
- You're the person other engineers come to when they want to bring AI into a workflow. You've built the foundation that lets them do it themselves.
Day-to-day- Sit with users. You'll spend hours in the workflows of sales, success, ops, finance, and engineering. Not because you have to. Because that's where the problems are.
- Ship narrow agents. Tight scope, real users, fast iteration. Whatever combination of agents, integrations, and internal tools the situation calls for.
- Operate our internal AI assistant as a product. Triage what's working, fill the gaps, follow up on questions that didn't get good answers, push the loop until quality is consistently high.
- Raise the engineering productivity bar. Agent-driven testing, review, and verification as part of the everyday loop. Automated migration tooling that compresses weeks into hours.
- Build the patterns and primitives that let other engineers do this kind of work themselves: skills, agent templates, MCP servers, eval harnesses.
- Measure adoption honestly. Build the dashboards that tell us who's actually using AI, where, and what the leverage is. Use them to make the next set of bets.
- Drive adoption. Some of the work is technical, some is human. You can sit with someone, watch them try a tool, figure out whether the gap is in the workflow or in the tool itself, and close it.
What we're looking for- Systems thinker, operationally focused. You see organizations as systems and have the persistence to keep going at a narrow problem until it's actually gone. If you've ever quietly built an agent that replaced a manual process and watched a team unconsciously absorb the new way of working, this is your job.
- Owner, not renter. You operate with agency. You take ambiguous problems and turn them into shipped outcomes without being handed a spec. You know when to push and when to stop, and have the discipline to kill what isn't working rather than defend it.
- Strategic, even in the embedded work. Day-to-day is internal: sit with teams, ship narrow agents, raise the productivity bar. But you hold the longer arc at the same time. Where AI could touch firm-facing surfaces, what new model capabilities unlock, where the team's mandate should grow next. We need someone who can do the embedded work without losing sight of the bigger picture.
- Built custom agents and AI tooling in the wild. You've shipped agents real people use, on whatever stack. You can describe one in detail: who used it, what changed, what broke, what you learned.
- AI-native fluency, deeply. Agent frameworks, MCP, eval harnesses, retrieval, prompt engineering. These are tools you reach for, not topics you've read about. You can describe where AI helps and where it creates noise.
- Comfortable outside engineering. You can run a discovery session with a non-technical team, walk away with a clear picture of their day, and ship something useful inside a week. Empathy and curiosity are non-negotiable.
- Strong full-stack technical chops. You can stand up a service, design an integration, debug a production issue, and write code that doesn't need a lot of cleanup. You won't be living in a single codebase, so generalist instincts matter more than depth in any one stack.
- Bias to ship over bias to discuss. You'd rather have a working v1 in front of a user by Friday than a perfect spec on Monday. You go from nothing to working in days, not sprints, and you know which version is good enough.
- Legibility. You can write up what you did, why, and what's next in a way anyone in the company can read and understand. This role is high-leverage, but only if it's visible.
- Bonus: Internal-tools or forward-deployed engineering background. DevEx, developer productivity, or platform engineering experience. B2B SaaS in a regulated or domain-heavy industry. History of building "Ops as a system" instead of scaling ops headcount. Public writing, talks, or open-source contributions in the developer-tooling-meets-AI space.
What we offer- The chance to define what AI-native work looks like inside one of the most consequential vertical SaaS companies of the next decade.
- A direct, reflective leadership team that actively welcomes challenge and debate.
- Real autonomy. You'll have budget, latitude, and the support to walk into any team and reshape how they work.
- Competitive compensation and meaningful equity.
Interview processWe'll share a detailed interview plan with you on the intro call. The loop is designed to mirror the actual job: real artifacts to react to, conversations that look like the conversations you'd have on the team, and reference checks taken seriously. We'll move quickly when there's mutual fit.
