Green Thumb Industries is building a data science function that powers real operational decisions — demand forecasting that drives inventory positioning, analytics science that surfaces what's happening in our stores, and feature engineering that makes every model smarter over time. This is a hands-on individual contributor role on a small, high-output, high-visibility team. You will spend your time building, testing, and maintaining ML models, engineering features, and translating data into answers that the business can act on. You will work closely with the Manager of Data Engineering, AI & ML, who will guide your technical direction and business context while you grow into shaping both. The systems are already starting to get built — your job is to push them further. This is a hybrid role and requires in office work 1 day per week every 2 weeks at our office in River North in downtown Chicago.
Responsibilities
ML Forecasting
Build, validate, and refine demand forecasting models for GTI's retail, wholesale, and other emerging business verticals across daily, weekly, monthly, and quarterly forecast horizons
Engineer new features for the Snowflake Feature Store — drawing from retail sales history, inventory movement, weather data, customer demographics, and external signals — to improve model accuracy across store, product, market and other dimensions
Develop and test new model candidates against GTI's established backtesting framework; interpret backtest results and surface findings to inform promotion decisions
Investigate forecasting errors and anomalies: identify when model performance degrades, diagnose root causes (data drift, structural breaks, new store openings, regulatory changes), and propose remediation
Conduct dimensionality reduction and principal component analysis to understand primary feature importance
Collaborate with the Manager to evolve the feature engineering roadmap — identifying signals worth building, data gaps worth closing, and model architectures worth exploring
Analytics Science
Design, validate, and execute analytical studies that answer business-user’s operational questions which can then be modeled and replicated by our data analyst AI agent to further promote self-service
Build reusable analytical frameworks on top of GTI's curated data layer (retail sales, inventory, customer, loyalty, workforce) that can be repeated, parameterized, and handed off to the business
Contribute to quasi-experimental modeling: pre/post adult-use launch performance, store cohort comparisons, product mix attribution, and discount effectiveness
Translate analytical findings into clear written summaries and visualizations that non-technical stakeholders can act on
Identify patterns in the data that surface new questions worth asking — and bring those to strategy discussions with the Manager
Collaboration & Growth
Participate in team roadmap and design discussions; contribute your analytical perspective on what problems are worth solving and how
Learn GTI's production data stack (Snowflake, dbt, Dagster) and the curated data models that underpin all analytical work — these are your primary data surfaces
Over time, develop familiarity with GTI's Snowflake based AI agent ecosystem and how structured analytical outputs feed into natural language intelligence tooling
Qualifications
2+ years of hands-on experience in a data science, quantitative analyst, or ML engineering role — with demonstrable work in model building, feature engineering, or statistical analysis
Strong Python skills for data manipulation, modeling, and analysis (pandas, scikit-learn, statsmodels, or equivalent). Jupyter notebook development or equivalent experience
Strong SQL skills — comfortable writing complex queries across multiple joined tables, aggregating at multiple grains, and debugging data quality issues in query output, while validating accuracy and trust
Working experience with supervised and unsupervised ML methods: gradient boosting, time series models, random forest, decision trees, etc
Ability to communicate analytical findings clearly in writing — you don't just run the analysis, you explain what it means and what to do about it
Intellectual curiosity and a bias toward figuring things out — this role requires navigating real, messy data in a complex multi-state retail operation
Preferred
Experience with time series forecasting methodologies (ARIMA, Prophet, LightGBM/XGBoost for tabular time series, or similar)
Experience with advanced machine learning modeling techniques and algorithms such as Bayesian inference, Deep Learning neural networks, k-means clustering, etc
Familiarity with feature store concepts or structured feature engineering pipelines
Exposure to Snowflake, Snowpark, or cloud data warehouse environments
Experience with dbt or working in a layered data warehouse (raw → refined → curated) — understanding where data comes from matters here
Experience prototyping and productionizing data products such as Streamlit apps
Basic familiarity with LLM-powered tooling or AI agent frameworks — not required, but exposure gives you context for where the team is headed
Background in retail, CPG, consumer analytics, or any multi-location operations business
Additional Requirements
Must pass any and all required background checks
Must be and remain compliant with all legal or company regulations for working in the industry
Must be a minimum of 21 years of age
#LI-HYBRID
The pay range is competitive and based on experience, qualifications, and/or location of the role. Positions may be eligible for a discretionary annual incentive program driven by organization and individual performance.
