New opportunity

Data Scientist

Mediacom Communications Corporation · Blooming Grove, United States

About this role

Mediacom Communications seeks a Data Scientist to develop and deploy data-driven, machine learning, and AI solutions for customer experience, retention, advertising effectiveness, and business performance. Working with marketing, sales, networking, and leadership teams, the role owns solutions from data discovery and design through deployment, monitoring, and optimization. The position is listed in Blooming Grove, NY, with a hybrid work designation.

Responsibilities

Gather requirements, write functional design documents, and analyze structured and unstructured data for trends and patterns.
Build predictive models to reduce customer churn, improve customer experience, and optimize ad targeting. Design, prototype, and support generative AI and agent-based solutions using LangChain and LangGraph, including retrieval-augmented generation, tool/function calling, workflow orchestration, API integration, and enterprise data connectivity.
Develop scalable data pipelines, analytics workflows, and machine learning and AI solutions in Databricks. Monitor model performance and data accuracy; implement AI observability and feedback loops.
Coordinate with business users, technical and outsourced teams, data engineers, developers, and external partners. Support production AI and analytics platforms, investigate issues, and implement mitigations.

Requirements

Bachelor’s degree in Computer Science, Statistics, Mathematics, Data Science, or a related field and at least 3 years of hands-on experience in AI engineering, machine learning engineering, data science, or a related role.
Experience implementing advanced machine learning algorithms and AI solutions using statistical analysis; knowledge of deep learning, model evaluation, and MLOps. Strong Python, SQL, and database skills; hands-on Databricks, notebook development, Spark-based processing, and machine learning workflow support.
The overview calls for experience with machine learning, LLMs, RAG, AI agents, prompt engineering, and production tuning. Understanding of AI governance, security, responsible AI, and model monitoring is required. Strong communication, project management, stakeholder relationship-building, and independent working skills are also required.

Preferred

A relevant master’s degree; hands-on development of AI agents using LangChain, LangGraph, LLMs, RAG, vector databases, APIs, and enterprise data sources; and cloud AI deployment experience with Azure, AWS, or Google Cloud.

Compensation and benefits: Salary range of $100,000–$135,000. Benefits include health, vision, and dental coverage; up to $5,000 annually for professional development and education; a matched 401(k); paid time off; training; and service discounts or internet/cable reimbursement where Mediacom services are unavailable.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

PythonSQLDatabricksApache SparkMachine LearningDeep LearningPredictive ModelingStatistical AnalysisModel EvaluationMLOpsData PipelinesData AnalysisGenerative AILarge Language ModelsRetrieval-Augmented GenerationAI AgentsLangChainLangGraphPrompt EngineeringVector DatabasesAPI IntegrationModel MonitoringAI ObservabilityAI GovernanceResponsible AIAI SecurityAzureAWSGoogle CloudRequirements Gathering and Analysis

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Predictive models tied to business outcomes
The role uses predictive modeling to improve customer experience, reduce churn, and optimize ad targeting. Model evaluation matters alongside the business outcome each model is intended to support.
Generative AI and agentic solutions
The posting calls for designing, prototyping, and supporting solutions using LangChain and LangGraph, including RAG, tool/function calling, workflow orchestration, APIs, and enterprise data connectivity.
Scalable delivery in Databricks
The role develops and maintains data pipelines, analytics workflows, and machine learning and AI solutions in Databricks. Hands-on notebook development, Spark processing, Python, and SQL are stated requirements.
Production monitoring and support
Ownership extends through deployment, monitoring, and optimization. Duties include checking model performance and data accuracy, implementing AI observability and feedback loops, and diagnosing and mitigating production issues.
Cross-functional solution ownership
The role gathers requirements, writes functional design documents, and coordinates with marketing, sales, networking, leadership, technical teams, outsourced team members, and external partners.

Likely challenges to think through

Balancing model performance with business usefulness
Preparation hypothesis, not a confirmed problem: a churn or ad-targeting model could score well yet be difficult to use effectively. Prepare to explain how you would define an outcome with stakeholders, select evaluation measures, and revisit results after deployment.
Making agent workflows dependable
Preparation hypothesis, not a confirmed problem: RAG, tool calls, and enterprise data access add capability but also create reliability, security, and governance trade-offs. Prepare an approach to testing outputs, constraining access, monitoring behavior, and using feedback to improve the solution.
Managing delivery across teams and production demands
Preparation hypothesis, not a confirmed problem: simultaneous projects and production support may compete for time while business and technical teams need different levels of detail. Prepare to describe how you would document decisions, prioritize a mitigation against planned work, and keep stakeholders aligned.

Practice your answers

Technical practice

How would you turn a business request into a functional design for an AI solution?

What to coverThe posting calls for requirements gathering, functional design documents, and ownership from data discovery through deployment. Suggested preparation: explain how you would define the business outcome, data needs, acceptance criteria, dependencies, and monitoring plan before selecting an approach.

How would you investigate patterns across structured and unstructured data?

What to coverThe role involves processing both data types and requires strong Python, SQL, and database skills. Suggested preparation: describe how you would assess data quality, join relevant records, extract usable signals from unstructured content, and check whether apparent patterns are reliable.

How would you build and evaluate a model intended to reduce customer churn?

What to coverPredictive modeling for customer retention is an explicit responsibility, and the posting calls for statistical analysis, machine learning, deep learning knowledge, and model evaluation. Suggested preparation: discuss target definition, leakage, validation, suitable metrics, and when a more complex model would justify its operational cost.

How would you assess whether a predictive model improves ad targeting?

What to coverThe role aims to improve advertising effectiveness and optimize ad targeting. Suggested preparation: distinguish predictive performance from business impact, explain how you would compare against a baseline, and address measurement bias or trade-offs between reach and relevance.

How would you design a scalable analytics and ML workflow in Databricks?

What to coverThe posting specifies Databricks pipelines and workflows, notebook development, and Spark-based processing. Suggested preparation: outline ingestion, transformations, data-quality checks, reproducible model runs, and how you would diagnose a slow or unreliable Spark stage.

What would you monitor after deploying a machine learning model?

What to coverThe role owns deployment and optimization and calls for monitoring model performance and data accuracy using MLOps practices. Suggested preparation: distinguish data-quality failures from changes in model performance, then explain alerting, investigation, and criteria for updating or rolling back a model.

How would you design and evaluate a retrieval-augmented generation application?

What to coverThe overview requires LLM and RAG experience, while the posting names prompt engineering, enterprise data sources, and vector databases. Suggested preparation: explain retrieval and prompt choices, how you would test answer grounding and usefulness, and the trade-off between retrieval coverage and irrelevant context.

When would you use an AI agent, and how would you control its workflow?

What to coverThe responsibilities name LangChain and LangGraph agent solutions involving tool or function calling, orchestration, APIs, and enterprise connectivity. Suggested preparation: describe a bounded workflow, tool permissions, failure handling, and how you would decide whether an agent adds value over a simpler application.

How would you apply governance and security controls to a production AI solution?

What to coverThe requirements include AI governance, security, responsible AI, model monitoring, and prompt-engineering practices. Suggested preparation: address access to enterprise data, sensitive information, unsafe or unsupported outputs, and how monitoring and human feedback would inform improvements.

How would you diagnose a failure in a production AI or analytics platform?

What to coverProduction support, root-cause investigation, and mitigation are stated responsibilities. Suggested preparation: explain how you would narrow the fault across data, pipelines, models, APIs, and application behavior, then verify a mitigation and prevent recurrence.

Behavioural practice

Tell me about a time you reconciled different stakeholders’ goals for a data or AI project.

What to coverThe role works across marketing, sales, networking, and leadership while gathering requirements. Suggested preparation: use a real situation/action/result example showing how you clarified competing needs, agreed on a scope, and measured the outcome.

Describe a time you translated between business users and technical contributors.

What to coverThe posting explicitly calls for liaison work with business users, technical colleagues, and outsourced team members. Suggested preparation: give a real situation/action/result example that shows how you made requirements understandable to each group and resolved a misunderstanding.

Tell me about a time you managed several technical projects at once.

What to coverThe requirements call for managing multiple projects simultaneously and producing tangible results. Suggested preparation: use a real situation/action/result example explaining your prioritization, how you communicated trade-offs, and what was delivered.

Describe a time you moved a project forward with limited direction.

What to coverThe posting calls for self-sufficiency and independent work across the solution lifecycle. Suggested preparation: give a real situation/action/result example showing which decisions you made independently, when you sought input, and how you verified the result.

Tell me about a time feedback changed a deployed analytics or AI solution.

What to coverThe role includes observability, feedback loops, monitoring, and ongoing optimization. Suggested preparation: use a real situation/action/result example covering the feedback or signal, the change you made with collaborators, and its observed effect.

Sharpen your knowledge

Topics to prepare

Predictive modeling and evaluation
Suggested exercise: explain how you would select, evaluate, and tune a model for customer churn, including how statistical analysis, deep learning, and business outcomes affect your choices. Relevance: the posting calls for advanced machine learning, model evaluation, and predictive models for retention, customer experience, and ad targeting.
RAG and agent architecture
Suggested exercise: diagram how an LLM application could retrieve enterprise information, call a function, and route work through an agent workflow; identify where prompt engineering and production tuning matter. Relevance: the overview requires experience with LLMs, RAG, agents, and prompt engineering, while the responsibilities name retrieval, tool calling, and orchestration.
Model lifecycle and observability
Suggested exercise: trace a solution from data discovery through deployment, then describe how you would detect degraded predictions or inaccurate data and decide what to investigate. Relevance: the role owns deployment, monitoring, optimization, feedback loops, and production support.
AI governance and security
Suggested exercise: review a hypothetical enterprise-data RAG design for access-control, unsafe-output, and monitoring risks, and explain possible mitigations. Relevance: understanding AI governance, security, and responsible AI is explicitly required.
Requirements-to-design translation
Suggested exercise: turn a hypothetical request to reduce churn into a problem statement, data questions, success measures, and a functional-design outline. Relevance: the role gathers requirements, writes functional design documents, and works across business and technical teams.

Skills to revise

Python and SQL data analysis
Suggested exercise: write SQL to assemble a hypothetical customer-level analysis table, then use Python to inspect missing values and summarize trends. Relevance: excellent Python, SQL, and database skills are required, and the role analyzes structured and unstructured data.
Databricks notebook and Spark workflow
Suggested exercise: sketch or build a Databricks notebook that reads data, applies Spark transformations, checks data accuracy, and produces modeling inputs. Relevance: hands-on Databricks, notebook development, Spark processing, and scalable pipelines are specified.
Predictive-model implementation
Suggested exercise: build a small churn-model prototype, compare it with a baseline, and explain the evaluation and tuning decisions. Relevance: implementing machine learning solutions and using predictive modeling to reduce churn are central duties.
Agent-based prototype
Suggested exercise: prototype a limited RAG workflow with a retrieval step and tool call, then explain how LangChain or LangGraph would orchestrate it and connect to an API. Relevance: the responsibilities name these frameworks and capabilities; prior hands-on agent development is listed as preferred.
Production issue investigation
Suggested exercise: given a hypothetical drop in model performance, outline checks for input accuracy, model behavior, and AI feedback, then propose a mitigation. Relevance: the posting calls for model monitoring, AI observability, root-cause investigation, and production support.
Functional design communication
Suggested exercise: draft a one-page functional design for a hypothetical ad-targeting model and explain its assumptions and open questions to both a business user and a data engineer. Relevance: the role writes design documents and liaises with stakeholders and technical teams.

Your preparation checklist

Map the stated requirements to evidence
Prepare a one-page reference to relevant examples of Python, SQL, statistical analysis, Databricks, Spark, model evaluation and MLOps work. Note the posting’s degree and three-year experience requirements without assuming they are met.
Draft a functional design exercise
Using a hypothetical customer-retention request, write sample stakeholder questions, a problem statement, data needs, success measures and a short functional design. The role includes requirements gathering and design documents.
Rehearse a predictive-model walkthrough
If you have a relevant project, explain the data, modeling and statistical choices, evaluation, business outcome and limitations. Connect it to the posting’s churn, customer-experience or ad-targeting use cases without implying you worked on Mediacom data.
Show the path from data to deployment
Prepare a diagram or walkthrough of relevant notebook, SQL, Spark and Databricks pipeline work, including how data accuracy is checked and an ML workflow is supported. Clearly label any hypothetical components.
Sketch an agentic AI design
For a hypothetical enterprise-data use case, outline a RAG and agent workflow with retrieval, prompts, tool or API calls and orchestration. Be ready to explain where LangChain or LangGraph could fit; hands-on agent development is listed as preferred.
Prepare a production-support example
If available, bring an example of diagnosing a model or data issue, mitigating it and monitoring the result. Cover model performance, AI observability, feedback loops, security and responsible-AI considerations called out in the posting.
Practice a cross-functional delivery story
Prepare one concise example, if available, of aligning business users with engineers or other partners while managing competing work independently. Identify your decisions, communication and tangible result.
Write questions for the employer
Ask how this role prioritizes predictive modeling versus generative AI, what enterprise data and Databricks workflows it uses, how production monitoring and governance are handled, and how marketing, sales, networking and technical teams collaborate.

Availability last checked: Nov 4, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…