Principal Growth Engineer — Meta Ads & AI Methodology

We're building AI agents for performance marketing.

Plurio, the AI agent for media buying teams running $300K–$15M/month on Google, Meta, and TikTok. It connects to your internal backend data and your ad accounts, automates analysis and optimization, saves 50% of time, and boosts ROAS by 10–30%. Watch a five-minute demo.

Our agents already run $500M+/year in ad spend (see this inside a $20M+/year User Acquisition team). Since our public release in March, our customer base has grown 4× in 10–12 weeks.

Our goal – and the mission we'd build together – is the autonomous user acquisition factory that runs 10× better than anything before it.

The role

You own The Algorithm: the methodology for running Meta Ads, encoded as the deterministic rules and workflows the agent runs across the whole client portfolio. It's an individual-contributor seat, closest to the product.

For this role, "done" means

• Given the same creatives and landing pages / apps, our media-buying algorithm performs much better than manual buying, and we can prove it.

• We have an algorithm that runs the whole Meta buying end to end from a set of inputs: goals (target CPA/ROAS, budget), parameters, and creatives.

• We generate hypotheses to improve the algorithm, test them, and keep what works.

• We have skills to audit, implement, adapt, and control the algorithm as the workflows and rules for a specific account.

What the methodology covers

The methodology holds everything about running Meta end to end: account structure, the optimisation event, test launches, scaling, burnout, budget allocation, and the data that makes all of it work better.

It's an adaptive system. Part of the methodology is standard on every account. The rest depends on various factors. That means working out how to define it for a specific account based on historical data, how to test for it when you can't define it from history, and how to deal with too few conversions, forecasted purchases and LTV, or degrading data quality.

Autonomy grows over time. First, separate rules per client. Then a standard set with customization. Then autonomous runs under guardrails and audits. Eventually it improves itself daily, with review.

The full picture, and how we think about the role, is on the Principal Growth Engineer board in Miro.

What you'll do

• Own the single source of truth: one written methodology that everything else follows, plus the rules and workflows that encode it.

• Improve The Algorithm: define how to value the gain or loss from each action or inaction on an account, generate suggested changes to the algorithm that could improve results, and test them on historical data.

• Build the testing machine: design and run tests by hand with the agent, then a testing skill that speeds up every step of the testing cycle, then a testing loop that runs cycles autonomously and generates improvements for review.

• Make it implementable: an account audit skill that account managers and AI marketing engineers use to check each account against The Algorithm for differences and improvements. Implementation skills that create and adapt every workflow and rule for a new client. Plus the control skills, automations, and loops that keep it running.

Who we're looking for

We're looking for the top-1 expert on Meta methodology, ready to put it inside an agent.

Must-have

• Deep Meta methodology you can put into words. You know Meta from the inside: account architecture, creative testing, data analysis, thresholds, run at volume so you know where CPM breaks and saturation sets in. And you can lay it out fast, clearly enough to hand to a team, a client, and the product, and to convince a strong buyer in minutes.

• You've reworked how a channel runs, down to the mechanics. You've done it more than once, sharpening your approach against historical data each time. This way of working is the core of the role.

• You've built a repeatable system across a portfolio. You've turned what works into a standard playbook or ruleset applied across many accounts at once, one approach with per-account modifications, and improved it as you scaled.

• You've built decision logic past simple rules. Combined metrics and rule cascades at fine granularity, each priced by the expected value of acting or holding back.

• AI-native, you build it yourself. You build the pipelines and automation by hand (vibe-coding rules, wiring models) and turn your logic into deterministic workflows, backtested before they ship.

• Working English. Enough to pull methodology out of the best operators and UA leads worldwide. You hold a working conversation comfortably; native fluency isn't required.

Nice to have

• Hands-on buying background of your own. You've run Meta campaigns in the account yourself, so your rules are pressure-tested against real buying.

What you'll get

Top-tier data to build on. We built Elly Analytics: a full-funnel attribution platform that combines every source into one source of truth. Multi-touch across web and app, with marketing-mix modeling, made for complex products with long cycles and delayed conversions. Every client is on it (our platform or their own equivalent in-house), and the agent sits on top. You test and sharpen your methodology across a large real-spend dataset (every segment, industry, and funnel), building rules that hold portfolio-wide.

• A machine built around your methodology. AI marketing engineers run and implement it on real client accounts, and the agent executes it. Clients bring their own data and context. The product and data teams support it, and an expert community trades findings with you. Your methodology compounds across the whole portfolio, well beyond what one person could carry alone.

• Together we write the Meta Methodology Playbook. That's the plan we'd build with you: the channel's methodology, turned into a playbook and embedded in the agent. It's how Plurio's clients run Meta, and the reference other teams in the segment look to.

• You join a team that runs on agents. Every function works through them in a shared workspace (see how we work as an AI-native team). We show you how to use and build them day-to-day, and how to put your expertise inside.

• Direct work with founders and leadership team. Kirill Kasimskiy and Seva Ustinov: built and scaled a performance marketing agency to 100+ people, then a full-funnel marketing analytics platform, then rebuilt their own team around AI agents.

How to apply

Learn more about Plurio

• Watch a five-minute demo of Plurio

• See how this looks inside a $20M+/year UA team

• See how we work as an AI-native team