Product Manager / Analyst — Analytics & ML

Hire5; Lisboa; Híbrido; EUR 85 000 - 120 000; Tempo integral

Resumo da oferta

Maestra.io is hiring a hands-on product manager / analyst to own the data side of the platform: merchant-facing reporting, experimentation methodology, and ML-powered product recommendations. You will design metrics, prototypes, and semantic layers to deliver trustworthy numbers merchants rely on.

You will work directly with the founder-CEO and collaborate with product and R&D teams. In this role you’ll own analytics end to end, guide A/B tests across campaigns and flows, and lead the evaluation

Qualificações

• Strong SQL with window functions and performance tuning.
• A/B testing math and statistics fundamentals.
• Judgment on modern ML tooling and evaluation against baselines.
• Direct collaboration with non-technical end users to translate needs into shipped metrics.

Responsabilidades

• Own merchant-facing analytics and the semantic layer across product areas.
• Define and audit experimentation methodology and defaults for merchants.
• Own ML product recommendations, evaluations, and improvement roadmap.

Conhecimentos

SQL

A/B testing

ML evaluation

Non-technical stakeholder comms

Clear written thinking

English & Russian

Ferramentas

Delta Lake

Trino

Spark

Descrição da oferta de emprego

About

Maestra.io is an all-in-one ecommerce marketing platform: a real-time customer data platform with email, SMS, push, on-site and product personalization, loyalty, referrals and product recommendations in a single system — sold with a dedicated marketer attached to every account. Mid-market DTC brands use it to replace four or five separate tools; Urban Armor Gear, 4ocean, Enlightened Equipment, Selkirk Sport and Bokksu are among them.

US ARR grew five-fold in 2025 and the company is approaching $5M ARR. Maestra is founder-funded, around 40 people, and building a long-term profitable business on the US market rather than optimizing for the next round.

The role

We’re hiring a hands-on product manager / analyst to own the data side of the platform: merchant-facing reporting, experimentation methodology, and ML-powered product recommendations.

Everything Maestra sells rests on numbers merchants trust. A brand consolidating five tools into one has to believe our attribution, our loyalty economics, our A/B test results and our recommendation performance — often when those numbers disagree with what they saw somewhere else. Today nobody owns that end to end. You would.

There is no dedicated data science team. You own the math and the experiments; R&D implements them in production. You report directly to the founder-CEO, who acts as CPO, with regular 1:1s, and another product manager will help you onboard.

What you own

• Merchant-facing analytics and the semantic layer. You design the metrics and reports across every product area — subscriber source attribution, loyalty program economics, recommendation performance, popup funnels, segment comparison. You take a report from raw-data reality to shipped product: spec the semantic layer (metrics, slices, aggregates, partitions), prototype it in SQL, hand it to engineering. And you untangle the methodology dead ends — cross-brand attribution, machine versus human email opens, multi-currency revenue, “why does this report disagree with the admin panel.”




• Experimentation methodology. You own how A/B testing works across the platform: widget tests, campaign tests, flow control groups. You audit them for correctness — comparable groups, honest participant counting, sample sizes, significance — and you define what “statistically correct out of the box” has to mean for hundreds of merchants who are not statisticians.




• ML product recommendations. You own our recommendation stack — 14 algorithms, including matrix factorization in Python — as a product: quality evaluation, degradation monitoring, debugging “this widget makes no money,” and the improvement roadmap. You also drive the paradigm call. Today’s models need dense per-product statistics; rare items and no-history customers fall back to bestsellers. You decide where modern embeddings and LLM re-ranking augment or replace classical collaborative filtering, and you prove it with honest evals against a dumb baseline.

How you’ll work

• AI-first. We expect heavy use of AI for SQL, Python, ML and, well, everything. AI writes — you verify. The statistics, the experimental design and the data model live in your head, because you are the last line of defense before a number reaches a customer.




• Deep questions over big teams. You’ll interrogate engineers, CSMs and merchants until the domain actually makes sense, and become the person everyone asks how a given number is computed.




• Tools. SQL over a multi-tenant lakehouse (Delta Lake / Trino / Spark, billions of rows), Python for analysis, Metabase for dashboard prototypes, Notion requirement cards, Slack threads with fast decisions and no bureaucracy.




• The team. Around 40 people in total; product and R&D is ten engineers, three product managers and the founder. You work horizontally with customer success and the other product managers rather than through layers — decisions happen in days, not quarters, and what you ship reaches real merchants within weeks.

What we’re looking for

• Strong SQL. Window functions over huge partitioned tables don’t scare you, and you understand SQL performance well enough to work with R&D on complex cases.




• A/B testing math and statistics fundamentals. Power and MDE, selection and survivorship bias, why you can’t condition on post-treatment variables — and how to turn that into defaults that protect merchants who will never read a statistics textbook.




• Judgment on the modern ML toolbox. Embeddings and vector similarity, when an LLM re-ranker helps and when it’s expensive noise, and how to evaluate any of it against a dumb baseline. You don’t need to have trained recommender models — you need to be able to judge them.




• Direct work with non-technical end users. You’ve sat with the people who actually use a product, uncovered what they were really trying to do, and turned it into something shipped. B2B product, consulting, solutions engineering and senior customer-facing roles all produce this.




• Clear written thinking. Much of the work — and our hiring process — is asynchronous. You can make a methodology argument in writing and have it land.




• Languages and hours. Russian (the working language inside the team) and English at B2+, fluency preferred. Based in the EU, with daily overlap with US Eastern time: R&D is mostly in the EU, customers and GTM are on the US East Coast.

A strong plus: ecommerce or martech domain experience; prior ownership of a metric or semantic layer; hands‑on exposure to recommender or ranking systems.

What this role is not

There is no analytics team to inherit and none to manage — this is a senior individual contributor role, and the three areas above currently have no owner. Maestra is founder‑funded and not a VC‑track business: skill growth here is close to guaranteed, while career and compensation growth track the company’s growth rather than a funding timeline. If you want a mature data platform to run and a defined ladder to climb, this isn’t that role. If you want to define how a fast‑growing product measures itself, it is.