Applied Data Scientist

Sterling Site Access Solutions LLC•Phoenix, IL

11h

About The Position

Before Sterling can prove anything, the data must be trustworthy. That's where this role starts. Sterling is building its data science function from scratch: greenfield infrastructure, a clear first problem (prove material changes extend mat life), and scope that grows into fleet intelligence, sales analytics, and technical diligence. You'll own the evidence stack needed to move initiatives through stage gates and build the models that turn messy field data into decisions the board will see.

Requirements

Bachelor's degree in Data Science, Statistics, Computer Science, Engineering, or similar quantitative discipline
3–10 years in data science, applied analytics, or ML engineering — ideally in a company with physical operations or industrial data
Demonstrated experience building, tuning, and deploying random forest and Monte Carlo models from scratch
Daily-driver Python (pandas, numpy, scipy, scikit-learn, statsmodels) and production-quality SQL against messy real-world schemas
Independent experience building ETL pipelines from raw sources to analysis-ready tables
Strong statistical reasoning: hypothesis testing, sample sizing, uncertainty quantification, and the judgment to distinguish directional signal from statistically significant conclusion
Track record of iterating models as new data arrives — adding features, retraining, recalibrating, or swapping algorithms
Experience building operational dashboards for non-technical stakeholders
Proven ability to communicate model outputs to non-technical audiences
Experience with messy, real-world datasets including sensor gaps, inconsistent labels, and small samples

Nice To Haves

Experience in construction, logistics, manufacturing, or industrial settings
Prior "one-person data science team" experience — built the pipeline, ran the model, presented the results without engineering support
Familiarity with CRM data (MS, Salesforce, HubSpot) for pipeline analytics or win-rate modeling
Financial modeling in code (NPV, DCF, business case) and ROI analysis
Exposure to RFID or IoT sensor data and field collection realities

Responsibilities

Audit and reconcile production records against system data across mats, RFID, and field collection sources. Establish 100% verifiability — every data point traceable, every discrepancy resolved — and maintain the lineage and quality standards that make downstream analysis defensible.
Build, tune, and maintain production-grade models — random forest, gradient boosting, regression, Monte Carlo — that quantify mat longevity, treatment efficacy, and the business value of material changes. Then iterate as new data arrives.
Maintain the confidence scoring methodology and reporting standards that gate every initiative before phase review. Ensure every claim — for IP, marketing, purchasing, or customers — is backed by appropriately sized, appropriately interpreted data.
Build repeatable pipelines from raw field, sensor, and RFID sources to clean, analysis-ready datasets. Operate independently and establish the documentation standards Sterling will scale on.
Turn raw RFID read data into fleet utilization, loss detection, and performance intelligence. Identify coverage gaps and reconcile reads against the physical fleet.
Partner with IT to build operational dashboards in PowerBI that non-technical stakeholders use. Translate model outputs into clear, honest communication for operators, executives, and customers.
As the foundation proves out, expand into CRM and pipeline analytics, win-rate modeling, marketing attribution, margin analysis, and the business cases that justify initiatives and capital allocation.
Build data cases for acquisition target evaluation — fleet quality, customer concentration, margin defensibility, and operational KPIs.
Define the data science standards, model documentation, and analytical best practices Sterling will scale with. Build the data function needed to power growth.