Product Data Scientist

About the role

Quarry is a boutique technology services firm, and we’re looking for a Product Data Scientist who is equal parts data scientist, data engineer, and technical product owner. You will sit on client account teams as the person who turns a client’s data into product decisions, and turns those decisions into work our engineers can ship.

This is not an internal analytics seat. Our clients retain Quarry, and you will work inside their product organizations as an outside partner: framing the question, doing the analysis end to end, telling the client what the data actually says (including when it contradicts what they hoped), and owning the backlog that follows. You will be in the room with directors, VPs, and founders, and your judgment about what to build, what not to build, and what it will cost is a large part of what the client is paying for.

AI is central to how we work. We expect you to use agents and models aggressively across the data and product lifecycle, and we expect the judgment behind every recommendation to be unmistakably yours.

Core responsibilities

• Analysis end to end: Take a client question from raw extracts to a defensible answer. Pull, clean, model, and validate data yourself; find the confounds before the client’s leadership does; write the finding in language a VP can act on.
• Data engineering: Build and maintain the pipelines, models, and instrumentation your analysis depends on when the client has none. You should be as comfortable writing an ELT job or an event taxonomy as you are fitting a model.
• Product ownership: Translate findings into a quarter’s worth of prioritized, sized work: problem statements, acceptance criteria, success metrics, and the explicit list of what we are not doing. Groom the backlog our lead engineer and designer pull from.
• Client delivery: Own the relationship on your workstream. Run the check-ins, write the status, deliver unwelcome news early and well, manage scope against the retainer, and recognize a change order when you see one.
• Measurement strategy: Help clients decide what to measure and why, from event design and experiment setup through to the dashboards and reviews that keep a product team honest.
• AI-native practice: Use agents for exploration, instrumentation, QA, and documentation, and set the standard on the team for doing so with rigor, reproducibility, and appropriate data handling.

What we’re looking for

We want someone who has done real analytical work with real stakes, has owned a software product roadmap, and can hold a client’s trust while telling them something they did not want to hear.

Technical mastery

• Applied data science: Strong statistical judgment, particularly around causal reasoning, confounding, cohorting, and the difference between what a dataset can and cannot support. Fluent in Python or R, SQL, and the modern notebook and warehouse toolchain.
• Data engineering fundamentals: Experience building pipelines, data models, and event instrumentation that other people rely on. Comfort with dbt-style modeling, orchestration, and API-based extraction where connectors do not exist.
• Product analytics: Retention, engagement, funnel, and cohort analysis; experiment design; subscription and renewal metrics.
• Product ownership: Prior accountability for a software product’s backlog and outcomes, including writing specifications engineers actually used.
• AI fluency: Demonstrated, current use of LLMs and agents in analytical and product work, with a clear view of where they are reliable and where they are not.

Professional attributes

• Client presence: You can run a meeting with a director or VP, disagree with them constructively, and leave them better positioned than you found them.
• Judgment under constraint: You size work honestly against a fixed budget, cut explicitly, and write down what you cut and why.
• Ownership: You do not wait for the question to be well-formed. You identify what the client should be asking and bring it to them.
• Writing: Short, calibrated, plain-language findings and proposals. You know what to leave out.
• Integrity with data: You handle customer data as the client’s customers would expect, and you say no to requests that would compromise that, however senior the requester.

Bonus points

• A scientific research background: you have designed studies, lived with messy real-world data, and defended results to skeptical reviewers.
• Prior time in a services, consultancy, or agency environment managing multiple client accounts.
• Experience on the client side of an agency relationship, and opinions about what the good ones did differently.

Why you’ll love working with us

• 100% remote & flexible: Work where you want, when you want. We value outcomes over hours.
• High-leverage impact: You will be the analytical and product voice on client engagements where the decisions are real and the audience is senior.
• AI-forward practice: We invest in agentic tooling and expect you to push it further.
• Competitive compensation: Top-of-market hourly rates for top-tier expertise.
• Professional growth: Stipends for equipment, continued education, and wellness.

How to apply

Please follow this application process carefully. Applications that include the Loom video get our full attention.

1. Instead of a traditional cover letter, record a short Loom video (max 5 minutes) with your camera on. Introduce yourself and answer the following three questions:
   • Question 1 — The unwelcome finding: Tell us about a time your analysis contradicted what a senior stakeholder had already decided. What did the data show, how did you deliver it, and what happened next?
   • Question 2 — From finding to backlog: Walk us through a product decision you owned that started with data. How did you size it, what did you choose not to do, and how did you know it worked?
   • Question 3 — AI in your practice: How are you using AI agents and models in analytical and product work today? Where have they changed your workflow, and where do you still not trust them?
2. Apply using the form below. Paste your Loom share link into the Loom intro field.
3. In the first line of Tell us about yourself, please write: “The data says no.”