Sr. Data Scientist

About Mariana Minerals

Mariana Minerals is a software-first, vertically integrated minerals company on a mission to supply the critical minerals powering modern energy, AI, and defense technologies. We’re reimagining the minerals supply chain by combining deep industry expertise with advanced software, automation, and data-driven decision-making.

Mariana Minerals is a software-first, vertically integrated minerals company supplying the minerals critical to modern energy, AI, and defense technologies. Our data doesn't live in a vacuum — it comes off sensors, assays, lab benches, and plant logs, and it's what the business uses to decide what to run, what to buy, what to fix, and what to build next.

We're hiring a Senior Data Scientist to be the person who makes those numbers trustworthy and legible. You'll own the statistics: how we measure recovery and throughput, how we know a process change actually worked, how much of a swing is signal and how much is assay noise, and what the reporting layer says when an operator or an executive opens it.

This is not an ML modeling role and not a data engineering role. Our MLEs build the models and our data engineers move the data. You bring statistical rigor and clear reporting to a company generating more data than it currently knows how to interpret, and you make sure the conclusions people draw from it are the right ones.

• Design and analyze plant trials and experiments — DOE, sample sizing, control selection, and the analysis that says whether a process change did what it was supposed to do and at what confidence.

• Own the definitions of the metrics the business runs on: recovery, grade, throughput, yield, uptime, unit cost. Decide what each one means, make the definition consistent across teams, and defend it when someone wants to compute it differently.

• Build and own the reporting and analytics layer — the recurring reporting, the dashboards, and the self-serve tooling that lets operators, engineers, and business leads answer their own questions.

• Quantify uncertainty honestly: sampling error, assay variability, instrument drift, measurement system analysis.

• Apply statistical process control and capability analysis to plant operations — know when a process has actually shifted versus when it’s a normal excursion.

• Run the deep-dive analyses that don’t have a home: why last month’s recovery dropped, what’s driving cost variance, which of these three suppliers is actually better, whether this correlation is real.

• Do forecasting and estimation for production, cost, and capacity planning — applied statistics that informs commitments, not research models.

• Partner with the Technical Product Manager for Data & Analytics Platform on what the reporting and analytics stack needs next, and be a demanding internal customer of the data platform when the data isn’t fit for purpose.

• Raise the analytical bar across the company: review other people’s analyses, catch the flawed comparison before it reaches a decision meeting, and teach the teams around you enough statistics to stop making the same mistake twice.

• Write findings up so they’re actually used — a clear memo a non-statistician can act on, not a notebook that needs you in the room.

• Rigor with a deadline: You know the difference between the analysis worth another week and the one that’s good enough to decide on today, and you say which is which.

• Skeptical by default: You ask how the data was collected before you analyze it, and you’re the person who notices the sensor was miscalibrated for three weeks.

• Structure from ambiguity: Turn a vague business question into a well-posed statistical one, get the data you need, and own the answer end to end.

• Explicit about confidence: Communi…