Senior Data Scientist — Data Cloud Acceleration

WHO WE ARE

Zeta Global (NYSE: ZETA) is the AI-Powered Marketing Cloud that leverages advanced artificial intelligence (AI) and trillions of consumer signals to make it easier for marketers to acquire, grow, and retain customers more efficiently. Through the Zeta Marketing Platform (ZMP), our vision is to make sophisticated marketing simple by unifying identity, intelligence, and omnichannel activation into a single platform – powered by one of the industry’s largest proprietary databases and AI.

Our enterprise customers across multiple verticals are empowered to personalize experiences with consumers at an individual level across every channel, delivering better results for marketing programs. Zeta was founded in 2007 by David A. Steinberg and John Sculley and is headquartered in New York City with offices around the world. To learn more, go to www.zetaglobal.com.

About the team

The Data Cloud Acceleration team identifies gaps and opportunities across clients and business units, then moves quickly to deliver practical new capabilities. We often develop and deploy the first version of a model, workflow, dataset, or application in days or weeks, learn from real usage, and improve it iteratively.

We are business-minded technologists who care more about impact than technical novelty. We use sophisticated methods when the problem requires them and simpler approaches when they will deliver a better result faster. Our work should be predictable, demoable, trusted, reusable, measured, and amplified by AI.

About the role

The Senior Data Scientist will build models, analyses, and supporting ML components that improve business decisions, intelligence products, and client outcomes. You will independently own defined deliverables—from understanding the requirement and preparing the data through modeling, validation, documentation, and delivery.

This is a hands-on individual contributor role. You will work across varied revenue and intelligence initiatives, often in partnership with a Lead Data Scientist, application engineers, analysts, and business stakeholders. The right candidate can move quickly without sacrificing trustworthiness and knows how to balance statistical rigor with the practical needs of the business.

What you’ll do

• Own model and analysis deliverables.
• Take a defined business problem and independently deliver a reliable model, analysis component, scoring workflow, or supporting dataset.
• Translate business questions into analytical approaches.
• Ask clarifying questions, understand how the output will be used, and recommend an approach that fits the decision, timeline, and available data.
• Build and test models quickly.
• Develop practical solutions using statistical methods, machine learning, deep learning, or existing models and services where appropriate.
• Prepare trustworthy data.
• Profile, cleanse, join, and validate noisy datasets while checking completeness, freshness, distributions, nulls, duplicates, and match rates.
• Create repeatable scoring workflows.
• Move useful work beyond the notebook by building reusable Python components, batch-scoring processes, APIs, or lightweight services.
• Evaluate results responsibly.
• Establish baselines, select appropriate metrics, perform statistical reasonableness checks, reconcile unexpected results, and clearly document limitations.
• Support intelligence products and applications.
• Work with application and data teams to define the right data ingredients, test hypotheses, and integrate model outputs into usable experiences.
• Add operational discipline.
• Include validation, monitoring, failure handling, refresh expectations, documentation, and a clear usage path in delivered work.
• Use AI to improve your own productivity.
• Apply tools such as Claude, Codex, and similar assistants to accelerate coding, testing, research, debugging, and documentation while independently verifying the results.
• Communicate progress early and clearly.
• Make milestones, assumptions, risks, dependencies, and issues visible rather than waiting until delivery.

What ownership looks like at this level

You are expected to own the deliverable

. That means:

• Working independently on a well-defined model, workflow, dataset, or component
• Producing reliable and repeatable outputs rather than one-time analyses
• Adding appropriate validation and basic monitoring
• Documenting assumptions, methodology, limitations, and usage
• Demonstrating the output and explaining how it supports the business
• Raising risks and ambiguity early
• Leaving the work in a condition that another team member can operate or extend

You will receive guidance on broader product direction, methodology, and complex stakeholder decisions, but you should not require step-by-step direction to complete the work.

Experience that can set you up for success

These are useful indicators rather than absolute requirements:

• Experience applying statistical analysis and machine learning to real business problems
• Strong Python skills and familiarity with libraries such as pandas, scikit-learn, XGBoost, LightGBM, PyTorch, or TensorFlow
• Strong SQL skills and experience analyzing large datasets
• Experience with cloud data platforms such as Snowflake, Databricks, Athena, Hive, BigQuery, or similar technologies
• Experience developing classification, regression, clustering, forecasting, recommendation, optimization, or anomaly-detection solutions
• Familiarity with model evaluation, experimental design, feature engineering, and statistical validation
• Experience creating repeatable batch-scoring workflows or exposing model outputs through APIs or services
• Familiarity with orchestration and automation tools such as Airflow, AWS Glue, Prefect, or similar platforms
• Experience using version control, testing, and reproducible development practices
• Ability to explain analytical results and trade-offs to technical and nontechnical stakeholders
• Meaningful use of GenAI tools to improve the speed and quality of day-to-day work
• You’ll thrive here if you

Are pragmatic.

You select the simplest credible approach that can deliver useful business impact.

Move quickly with discipline.

You can produce an initial version rapidly while still validating the fundamentals.

Care about trust.

You check the data, question surprising results, and make limitations visible.

Work well with ambiguity.

You can turn an incomplete request into a clear set of questions, assumptions, and next steps.

Think beyond the notebook.

You consider how a model will be refreshed, accessed, demonstrated, monitored, and reused.

Understand the business context.

You evaluate technical decisions through the lens of client outcomes, revenue, cost, adoption, and decision quality.

Are curious and low-ego.

You are comfortable learning from others, revising an approach, and using an existing solution when it is better than building a new one.

Early indicators of success

Within the first several months, a successful candidate will have:

• Independently delivered a model, analysis workflow, or scoring component used in an active business or intelligence initiative
• Added repeatable validation and documentation to their deliverables
• Converted at least one exploratory analysis or prototype into a reusable workflow
• Demonstrated clear understanding of how their work supports a client, product, revenue, or operational outcome
• Used AI-assisted development to improve delivery speed without compromising accuracy or trust
• Earned confidence from data science, application, and business partners through reliable execution and clear communication