Sr. Data Scientist

Mariana Minerals is looking for a Senior Data Scientist to make plant and business metrics trustworthy and legible for the people who run operations and the leaders who set direction. This role is grounded in applied statistics and uncertainty quantification, with reporting and analytics that operators, engineers, and business stakeholders can use day to day.

Location: San Francisco, CA (onsite)
Compensation Range: $140,000 - $173,000 per year

What you’ll focus on

• Design and analyze plant trials and experiments using DOE, including sample sizing, control selection, and the statistical conclusions that explain whether a process change worked and with what confidence.
• Own the definitions of business metrics including recovery, grade, throughput, yield, uptime, and unit cost, ensuring consistent interpretation across teams and being able to defend metric definitions when requested changes arise.
• Build and own the reporting and analytics layer: recurring reporting, dashboards, and self-serve tooling that helps operators, engineers, and business leads answer questions independently.
• Quantify uncertainty with honesty using sources such as sampling error, assay variability, instrument drift, and measurement system analysis.
• Apply statistical process control and capability analysis to identify when operations have truly shifted versus when results are within normal variation.
• Lead deep-dive analyses with no natural owner, such as explaining why recovery changed last month, identifying drivers of cost variance, comparing suppliers statistically, and assessing whether observed correlations are real.
• Perform forecasting and estimation for production, cost, and capacity planning, focusing on applied decisions rather than research modeling.
• Partner with the Technical Product Manager for the Data & Analytics Platform to shape what the reporting and analytics stack needs next, and act as a demanding internal customer when data is not fit for purpose.
• Raise the analytical bar through review and coaching: catch flawed comparisons before decision meetings and teach statistical fundamentals so repeated mistakes stop happening.
• Write findings in a way that leads to action, producing clear memos that a non-statistician can use without needing you present.

Requirements

• 3–6+ years in data science, statistics, analytics, or a quantitative research role where you owned analyses that drove real decisions.
• Strong applied statistics, including experimental design, hypothesis testing, regression, and uncertainty quantification, plus judgment about which tool fits the question and which assumptions may be violated.
• Strong SQL and Python (pandas, statsmodels, scipy) to work directly with data, run your own analysis, and produce your own reporting.
• A track record building reporting and dashboards people actually use, with the product sense to distinguish what belongs in a dashboard versus a memo.
• Comfort working with messy real-world measurement data, including missing values, inconsistent sampling, and instrument error, along with clarity on what the data can and cannot support.
• Exceptional written communication to translate technical findings for operators, engineers, and executives.
• Comfort being the statistical authority in the room.
• Background in mining, metallurgy, chemicals, energy, manufacturing, or other heavy industry, especially metallurgical accounting, mass balance reconciliation, or sampling theory.
• Experience with statistical process control or measurement system analysis.
• Experience with BI and analytics tooling (Hex, Looker, or embedded analytics) and an opinion on how to deploy them.
• Degree in statistics, chemical engineering, chemistry, operations research, economics, or a related quantitative field.
• Experience building an analytics function at a company where one did not exist yet.
• Working fluency with LLM-assisted analysis, with a clear understanding of where it speeds up work versus where it quietly introduces errors.

How you’ll operate

• Rigor with a deadline: know which analysis can be made “good enough to decide today” and when more time is warranted.
• Skeptical by default: question data collection before analyzing and notice measurement issues such as miscalibration.
• Structure from ambiguity: translate business questions into well-posed statistical problems and own the answer end to end.
• Explicit about confidence: communicate uncertainty in a decision-useful way.
• Ecosystem fluency: understand data origins, how downstream models and simulators use it, and how the reporting layer relates to the platform beneath it.

Company culture

• Everyone Gets Home Safe: never put speed or cost ahead of people.
• Extreme Ownership: take full responsibility for outcomes and drive toward solutions.
• Engineer Out Requirements, then Automate: simplify, optimize, and automate for scale.
• Share Your Legos: collaborate openly, share knowledge, and empower others.

Why this role

This position is explicitly focused on applied statistics, uncertainty quantification, and analytics that operators and executives can act on. It is not an ML modeling role or a data engineering role.

Technologies you’ll use

• SQL, Python (pandas, statsmodels, scipy)
• Hex, Looker
• LLM-assisted analysis