Senior Data Scientist

Zonos•St. George, UT

10h•$200,000 - $275,000•Onsite

About The Position

Zonos is seeking a data scientist to solve complex data problems that are integral to the company's AI operating system for cross-border trade. This role involves taking problems from conceptualization to production, working with diverse data types, and applying a range of statistical and machine learning techniques. The data scientist will report directly to the Chief Data Scientist and collaborate closely with engineering teams. The position is based on-site at the company's headquarters in St. George, Utah.

Requirements

5+ years of hands-on data science work with a Bachelor's degree, or 3+ years with a Master's or PhD, including things you built that ran in production.
Real depth in statistics: inference, experimental design, and uncertainty.
You understand what the numbers mean, not just how to produce them.
Practical experience with ML and LLMs/NLP, including fine-tuning, embeddings, evaluation, and text classification at scale.
Strong Python and SQL.
Ability to own the full loop: framing a problem, building the solution, shipping it, and staying with it after launch.
Ability to point to impact: shipped work that moved a real metric, and can explain how it was measured.
Write code engineers are happy to merge.
Lead with product judgment: ask 'should we solve this?' before 'how do we solve this?'

Nice To Haves

Optimization or operations research
Causal inference
Forecasting
Bayesian methods
Serving models on GPUs
Experience with trade, customs, logistics, ecommerce, or payments data.

Responsibilities

Own data problems at the heart of the product, such as classification, valuation, risk, and compliance.
Choose the approach, whether that's statistical modeling, classical ML, deep learning and LLMs, optimization, or a simple rule.
Work with messy, multilingual product and shipment data.
Figure out what it actually says before modeling anything.
Quantify uncertainty and make it useful.
Decide when we're confident enough to act, when to route work to a human, and what each error costs.
Design the evaluations, experiments, and analyses that tell us whether something actually worked, and hold the line when it didn't.
Write production-quality Python and SQL, then ship, monitor, and iterate on what you build.
Join the on-call rotation and debug in production, including the subtle failures that only show up on real traffic.
Bring rigor to product and business decisions beyond your own projects.