Senior Applied Data Scientist – Forecasting & Decision Systems

The role centres on designing and deploying forecasting, optimization, and AI-enabled decision systems to optimize Hawaii Foodservice Alliance’s logistics operations, with a hybrid work arrangement based in Oahu.

Responsibilities

• Lead the end-to-end demand forecasting capability from problem framing to production deployment, monitoring, and ongoing improvement.
• Develop store-item-day forecasts that capture seasonality, promotions, intermittent demand, stockouts, product life cycles, and the asymmetric costs of over or under ordering.
• Set up credible baselines, backtesting frameworks, business-weighted evaluation metrics, and controlled pilots.
• Rapidly prototype analytical applications, model-review tools, and operational workflows using modern AI-assisted development platforms.
• Translate model outputs into actionable recommendations for buyers, order writers, merchandisers, and operational leaders.
• Build production‑level Python and SQL solutions, including batch pipelines, model services, APIs, monitoring, and decision logs.
• Measure whether models produce meaningful business outcomes such as reduced distress, improved shelf availability, better ordering, and more efficient allocation of labor and resources.
• Design mechanisms to capture user overrides, operational context, and actual outcomes to enable continuous system improvement.
• Partner with Data Engineering and Business Intelligence to leverage and extend governed Snowflake and dbt data products.
• Collaborate with Software Engineering to integrate models and decision logic into HFA’s internal applications.
• Contribute to analytical datasets, KPI definitions, and dashboards that directly support the decision products you own.
• Explore optimization approaches for scheduling, routing, purchasing, replenishment, and task prioritization.
• Help develop AI-assisted operational monitoring that detects abnormalities, assembles supporting evidence, explains causes, and recommends next actions with human oversight.
• Clearly communicate findings to both technical teams and business leaders, including when evidence does not justify deploying a model.

Requirements

• At least five years of experience building applied data science, forecasting, machine learning, or decision-support systems, or an equivalent record of production ownership.
• Proven ability to move models from experimentation into live business workflows.
• Strong Python and SQL skills.
• Practical depth in time-series forecasting with rigorous backtesting and benchmarking against simple baselines.
• A scientific mindset grounded in hypothesis testing, measurement, uncertainty, and willingness to revise approaches based on evidence.
• Ability to translate ambiguous business problems into measurable objectives, constraints, evaluation criteria, and an implementable product.
• Experience delivering working software beyond notebooks, including applications, services, automated pipelines, or internal tools.
• Experience using AI-assisted coding tools to accelerate development while maintaining tests, version control, security, and maintainable architecture.
• Familiarity with production practices such as Docker, orchestration, CI/CD, model monitoring, drift detection, observability, or experiment tracking.
• Ability to work directly with operational users, understand decision processes, and explain technical results in practical terms.
• Strong judgment on when to apply advanced modeling versus simpler, more interpretable approaches.
• Ability to operate independently while collaborating closely with data, BI, software, and business teams.

Technologies

• Python, SQL, Snowflake, dbt, Snowpark, Airflow, MLflow, OR-Tools, Gurobi, CPLEX, Pyomo
• React, Streamlit, Dash
• Docker

Preferred Experience

• Demand forecasting in grocery, retail, consumer packaged goods, food distribution, replenishment, or other inventory-intensive settings
• Probabilistic or quantile forecasting, hierarchical forecasting, intermittent demand, cold-start forecasting, or demand censored by stockouts
• Inventory optimization, safety stock, service-level modeling, or newsvendor-type decisions
• Mathematical optimization including routing, scheduling, constraint programming, simulation, or mixed-integer programming
• Experience with OR-Tools, Gurobi, CPLEX, Pyomo or similar optimization frameworks
• Exposure to Snowflake, dbt, Airflow, MLflow, Snowpark or similar warehouse-centric platforms
• Applied LLM or agentic-system experience with tools, structured data, business rules, and human approval workflows
• Computer vision, OCR, or vision-language models
• Causal inference, promotion-lift analysis, difference-in-differences, synthetic controls, or experimentation in operational settings
• Experience building lightweight front ends or internal apps using React, Streamlit, or Dash
• Experience with DSD, logistics, warehouse, merchandising, transportation, or multi-location field operations

What Success Looks Like

• Develop a clear understanding of HFA’s ordering processes, constraints, and current forecasting methods
• Establish a reproducible backtesting and evaluation framework
• Replicate the incumbent baseline and identify key forecast error sources
• Deliver an initial model and forecast-review prototype for a focused category or area
• Show measurable improvement over the baseline on holdout data
• Deploy probabilistic forecasts or order recommendations in a controlled pilot
• Implement monitoring, drift detection, model versioning, and user-override logging
• Set up a plan to measure distress, shelf availability, service level, and adoption
• Assess operational and financial impact of the initial forecast pilot
• Extend successful approaches to additional categories or locations
• Establish a sustainable model development and monitoring process
• Deliver or scope a second decision system such as routing, scheduling, or AI-assisted monitoring

Compensation

• Salary Range: $145,000–$180,000 per year
• Actual compensation determined by qualifications, production experience, technical depth, and internal equity. This range reflects the base salary for the Oahu-based role at posting time.
• Candidates near the upper end typically bring production forecasting or optimization ownership, strong software development capabilities, and a track record of measurable operational impact.

Benefits & Wellness

• 100% employer-paid medical coverage for team members, with subsidized family coverage
• Dental, vision, and preventive care benefits
• 401(k) retirement plan with up to 4% employer match and immediate vesting
• Paid vacation, sick leave, holidays, and floating holidays
• Employee Assistance Program with free, confidential counseling
• Horizon Day — a paid annual day off to volunteer
• Wellness programs including gym membership discounts and health initiatives
• Flexible spending account
• Life insurance
• Paid time off

Why This Role

• You will directly own a new applied-science capability and shape systems that influence real orders, inventory, routes, labor, shelves, and customer service across Hawaiʻi.
• This is a hands-on role that blends scientific thinking with modern AI enabled development and strong engineering judgment to address meaningful operational problems, rather than notebook-only research or traditional reporting.