Senior Data Scientist

Roark is an Atlanta-based private equity firm with over $41 billion in assets under management. Roark focuses on investments in consumer and business service companies, with a specialization in franchise and franchise-like business models.

Roark prides itself on being a trusted partner for management and business owners. We believe that successful partnerships begin with providing a compelling value proposition to all business constituencies. A win‐win‐win approach leads to a positive business environment where all stakeholders – customers, employees, franchisees, strategic partners, service providers, management and shareholders – share in the growth and success of our businesses. We want to partner with entrepreneurs and executives who share this vision.

Roark brands generate approximately $100 billion in annual system revenues from 117,000+ locations located in 50 states and 120 countries.

Our portfolio companies span multiple industries including food, restaurants, consumer and business services, health, wellness and fitness, and education and youth activities. We are best known for our portfolio companies such as, Dunkin', Jimmy John’s, Jamba, Orangetheory, Mathnasium, Drybar just to name a few.

Roark is growing our in-house Data and Analytics team, with the goal of better leveraging advanced analytics, data science, and machine learning to make sound investment decisions and guide our owned businesses to profitable growth. We are seeking a Senior Data Scientist to ship production-grade models and tools across Roark investment use-cases and our brands. We are a fast-paced, growing, and entrepreneurial team, and this role will enjoy working across many topics and brands as we build durable competitive advantage in how we underwrite investments and scale portfolio performance.

This is a hands-on, senior individual-contributor role for a generalist data scientist with strong MLOps and software-engineering depth. Typical work spans regression modeling, demand forecasting, customer and unit economics modeling, pricing and promotion analytics, classification problems, and GenAI productivity tooling. You will own models end-to-end — from framing the problem through deployment, monitoring, and maintenance, in addition to taking existing models built by others and transforming them into to be more reusable and/or put into production.

We want a self-sufficient developer who raises the craft of the people around them. The team will look to this individual to help up-level their capabilities. The ideal candidate is comfortable across a wide variety of tasks, gets up to speed on a new project quickly, and adds value in an agile fashion. They work in an 80/20 style and do not need the perfect dataset to get started. This role partners closely with investment teams, our owned businesses, and Data & Software Engineering to deliver scalable, production-ready solutions that drive real business impact. This position is expected to be in-office to support collaboration and culture building.

• POSITION RESPONSIBILITIES
• Responsibilities include but are not limited to:

Own machine learning models end-to-end — build, deploy, monitor, and maintain them in production, including forecasting, regression, and classification, beyond proof-of-concept

Translate ambiguous business problems into well-defined modeling approaches and data requirements, working in an 80/20 style without waiting for a perfect dataset

Write clean, tested, reusable Python; participate in code review; and apply strong software craft, including Git and packaging

Deploy and operate models on the cloud (Google Cloud Platform preferred — Vertex AI, BigQuery, Cloud Run), applying MLOps fundamentals: CI/CD for ML, reproducibility, deployment, and monitoring

Raise the engineering and modeling standards of those around you — teaching craft and lifting team quality, not only your own output

Partner directly with investment teams, owned-business stakeholders, and internal clients to deliver analysis and tools that drive decisions

Author reusable IP — internal libraries, templates, and standards documentation — that scales across Roark’s businesses and target sectors

Get up to speed quickly on new projects and brands, adding value across a wide variety of tasks in an agile fashion

Ensure rigor in model validation, performance monitoring, and documentation, balancing sophistication with interpretability, speed, and impact

Stay current on advances in machine learning, AI, and applied analytics relevant to private equity and consumer businesses

POSITION QUALIFICATIONS

Bachelor’s degree in a quantitative discipline (computer science, data science, engineering, mathematics, statistics, economics, or similar), Master’s Degree preferred.

5–8 years of applied data science experience, operating with senior-level autonomy and requires minimal coaching or guidance

Demonstrated breadth: has shipped a variety of models; from forecasting, regression, and classification work, and is comfortable across the core algorithm families

Has owned at least one model in production — deployed, monitored, and maintained (not proof-of-concept only)

Strong software craft in Python: clean, tested, reusable code; Git; code review; and packaging

Cloud deployment experience (Google Cloud Platform preferred — Vertex AI, BigQuery, Cloud Run) and MLOps fundamentals you can teach others: CI/CD for ML, reproducibility, deployment, and monitoring

Evidence of having raised others’ engineering standards, not only your own

Client- and stakeholder-facing delivery experience, Either Consulting engagements or as cross-functional business support.

Applicants must be currently authorized to work in the United States without the need for visa sponsorship now or in the future; role is based in Atlanta, GA

Core areas of desired expertise include:

Regression and classification across a range of business problems

Pricing and promotion analytics

Marketing Analytics

Customer and unit economics modeling

Experiment design and measurement

Forecasting and demand prediction, including unit-level forecasting

Tools and platforms:

Python and SQL

Cloud platforms, Google Cloud Platform preferred (Vertex AI, BigQuery, Cloud Run)

MLOps tooling: CI/CD for ML, reproducibility, deployment, and monitoring

Generative AI coding productivity tools (Claude Code, Cursor, etc.) – Candidate should be comfortable using AI tools to be a more efficient and stronger Data Scientist.

Ideal candidates may also possess the following qualifications:

Generative AI / agentic development experience — Mastra, LangGraph, RAG, and LLM applications

Domain adjacency: QSR, multi-unit consumer, retail, CPG, or franchise businesses; unit-level forecasting

Orchestration and pipeline experience: Airflow, dbt, Dataflow

Has authored reusable IP — internal libraries, templates, or standards documentation

Desired personal and leadership characteristics include:

High standards of integrity, accountability, character, and professionalism

A self-sufficient Developer who delivers on time analysis and code

Raises the craft of those around them rather than only their own output

Hands-on, pragmatic, and results-oriented, with a bias for action and a willingness to roll up sleeves to get the job done

Comfortable working across a wide variety of tasks; gets up to speed on a new project quickly and adds value in an agile fashion

Works in an 80/20 style — does not need the perfect dataset to get started, and balances speed and impact with technical rigor

Thrives in a fast-paced, high-demand environment with multiple competing priorities; self-starter with a sense of urgency to deliver under tight deadlines

Low-ego team player who enjoys collaborating with high-intellect colleagues and stakeholders across various functions and backgrounds

Effective interpersonal and communication skills