AI Systems Lead

Sustain Media; Brasil; A distancia; BRL 281.000 - 393.000; Jornada completa

Descripción de la vacante

Sustain Media is seeking an AI Systems Lead to build and run the internal systems our team uses to deliver local SEO and Google Ads work. A big part of the role is turning ideas into builds and turning problems into reliable systems, including Viktor and our growing automation suite.

The role focuses on tuning Viktor, extending its capabilities, and owning the Airtable bases and integrations that power the team.

Responsabilidades

• Turn ideas into builds with the SEO, Google Ads and client success teams; create a clear plan before you build.
• Connect tools through Airtable, Slack, Google's ad and search tools; write custom code when needed; use n8n or Zapier when faster.
• Improve Viktor’s performance by tuning instructions and workflows, and selecting the right model for each task; build future agents with tests and reviews.
• Write down architectural choices and data structure decisions so future engineers understand them and avoid getting stuck.
• Own reliability: test before launch, handle missing data, set up canaries, logs, and alerts to catch failures early.
• Create simple documentation and handoff guides for non-technical managers; adjust workflows to fit the team.
• Keep projects moving with milestones, block blockers early, and clearly communicate what shipped and what’s next.

Descripción del empleo

Sustain Media | Remote | Full-time contractor | $5,000 to $7,000/month

Sustain Media helps home service businesses get more customers through local SEO and Google Ads. We're hiring an AI Systems Lead to build and run the internal systems our team uses to deliver that work.

A big part of this job is turning ideas into builds. Someone on the team describes a problem they deal with every week. You ask the right questions, whiteboard a solution with them, and turn it into a system the team can depend on.

We run a lot of the business off Viktor, an AI agent that lives in our Slack. It pulls data from our ad accounts and tools, runs SEO audits, tracks rankings and builds reports. A lot of work has already gone into Viktor and it runs well. Your job is to keep making it better and to build what comes next, including our own CRM and suite of SEO/Ads automations. You'll also own the systems around Viktor, including our Airtable bases. We need someone who can take a problem from the first conversation through launch and keep the system working after people start using it.

The kind of builder we want

You're practical. You'd rather ship a working version this week than spend a month on the perfect setup. But you don't leave a mess behind. You document what you built well enough that the team can use it and another engineer could pick it up.

You also think about where a system will be in a year. When you pick a tool or decide how data is structured, you can explain why, and you've checked that it won't box us in once we have twice the clients and twice the team. Choosing Airtable as our system of record is the kind of call we mean.

What you'll own

• Turn ideas into builds. Sit with our SEO, Google Ads and client success teams and learn how the work runs today. Take their user stories, whiteboard a solution with them, and turn a rough idea into a clear plan before you build. You're comfortable doing this live in front of a group. Tell us when something isn't worth automating.
• Connect our tools. Most of the work runs through Airtable, Slack, Google's ad and search tools, and LLM APIs. Write custom code when an existing connector falls short. Use n8n or Zapier when that's the faster call.
• Make Viktor perform way better. You don't need to build an agent from scratch (for now). Most of the work is tuning one that's already built: improve its instructions and workflows, connect it to the right data, and pick the right model for each job. When we do build new agents, you harden them with test cases, scheduled jobs and reviews so they stay on the path they're meant to follow.
• Make architecture decisions you can defend. Write down why you chose a tool, a data structure or a place for data to live, so the next person understands it and we don't get stuck later.
• Own reliability. Test before launch, handle missing data and duplicate events, and validate data on the back end, not just the front end. Set up canaries, logs and alerts so nothing fails without someone knowing, and fix failures before they become recurring problems.
• Make the systems usable. Write simple documentation and handoff guides so a non-technical manager can understand what you built and train the team on it. Adjust the workflow when it does not fit the team's day-to-day work.
• Keep projects moving. Set clear milestones, flag blockers early, and explain what shipped, what changed and what you're working on next.

Examples of the work

Some of this is already running and some is on the list: an onboarding portal for new clients, a Google Maps rank tracker, AI-drafted replies to Google reviews, a flow that hands tasks to the right person on the team, billing clients based on their Google Ads spend, and research that runs automatically before sales calls. You'll inherit some of it, improve it and build the rest. You'll help choose and scope new work based on its value to the business.

What we're looking for

• You've built automations that real teams rely on. You can show what you personally built, explain the decisions you made, and describe what happened after launch.
• You've worked on both the front end and the back end. You understand back-end validation and authentication, and you know how to make an AI-built app safe before it goes live. Ideally you were writing code before AI tools made it easy.
• You've built and managed databases that a team uses every day, in Airtable or other tools. You understand why different systems structure data the way they do, and you spot data problems that will break a workflow later.
• You've used LLMs inside working systems. You understand structured outputs, tool calling, model selection, testing, cost and the limits of model reliability.
• You can explain your choices. When you design a system, you can say why you built it that way, what you traded off and what would make you change it.
• You're easy to work with. You ask good questions before building, troubleshoot on your own, and can explain how a system works to people who aren't engineers. You're comfortable telling us when an idea won't work.
• Experience with marketing agencies, SEO, Google Ads, reporting or CRMs is incredibly useful. So is experience running or tuning your own models. We care most about the quality of your work and your ability to take ownership. A specific degree or AI framework is not required.

You're probably not a fit if

• You care more about code quality scores and server setups than whether the team's problem got solved.
• Your experience is mostly prompts and LLMs, and you haven't connected real apps through APIs.
• You'd rather get a finished spec than talk to the people doing the work.

What success looks like

When the team brings you an idea, it turns into a working system quickly. The team spends less time moving data, checking routine work and chasing information. They trust the systems because the outputs are accurate, failures are visible, and someone owns the fix. Your documentation makes it possible for another person to understand and maintain what you built.

Why this role

You'll work directly with the team and the people using your systems, and you'll see the effect of your work. You'll have room to recommend better ways of doing things and influence which problems we tackle next. This role grows with the company, with room for bonuses and a clear path to higher pay. You won't manage people or run operations. The job is building, with a lot of time spent talking to the people you build for.