Lead Data Scientist

Data Science Lead – Churn & Retention Analytics | Microsoft Fabric / Azure ML

Location: Gurugram, Haryana, India (overlap with US Central Time required)

Company: Neurogent.ai

Employment Type: Full-time (Hybrid)

Seniority: Senior / Lead

Experience: 7+ years

ABOUT NEUROGENT.AI:-

Neurogent.ai builds agentic AI solutions, including AI agents, chatbots, and voice assistants, alongside data and ML platforms that automate workflows and improve customer experience across banking, healthcare, insurance, financial services, and multi-location consumer businesses.

ABOUT THE ROLE:-

We are looking for a senior, hands-on Data Science Lead specializing in Churn & Retention Analytics to join our delivery team working with a leading US early-education and childcare provider serving approximately 130,000 children across a workforce of 20,000 staff.

The client is building its data science capability on Microsoft Fabric, with student and family churn as the first approved use case.

The objective is to identify families at risk of withdrawal 4–8 weeks before they leave, giving center directors enough time to intervene and improve retention.

Families often show early warning signs, such as declining attendance, reduced schedules, or changes in billing behavior. Your role will be to turn these signals into a reliable, explainable churn model that operations teams trust and use every week.

You will own the use case end-to-end, from defining churn and engineering features to building, deploying, and monitoring a production model that scores every active enrollment weekly.

This is both a technical leadership and business consulting role. You will work directly with US-based business leaders, the client's data science and IT teams, our onshore colleagues, and data engineers.

KEY RESPONSIBILITIES:-

1. Churn Analysis & Predictive Modeling

• Define churn with business stakeholders, including withdrawal criteria, observation and prediction windows, and preventable versus unavoidable exits.

• Engineer behavioral features from attendance, enrollment, schedule changes, tenure, sibling status, age relative to kindergarten eligibility, and payment or billing patterns.

• Build and validate tabular churn models using XGBoost, LightGBM, gradient boosting, survival analysis, or time-to-event modeling.

• Calibrate risk scores and generate actionable risk rankings at the center level.

• Use SHAP and other explainability techniques to communicate why a family is flagged and recommend appropriate retention actions.

2. Business Consulting & Stakeholder Management

• Collaborate with US-based operations leaders, regional teams, corporate data science, and executive stakeholders to define business problems and success metrics.

• Translate model outputs into clear business insights, retention playbooks, and revenue-impact estimates.

• Present findings and recommendations to senior stakeholders in a business-friendly manner.

• Challenge assumptions and requirements when necessary to ensure the solution drives measurable business outcomes.

3. Data Quality & Feature Engineering

• Partner with data engineering teams to integrate attendance, CRM, enrollment, and billing data into OneLake.

• Implement data-quality checks to identify source-system outages, missing attendance records, and other issues that could generate misleading churn signals.

• Establish safeguards to prevent model scoring when critical data-quality conditions are not met.

• Build reusable, versioned feature tables using Delta Lake to support future use cases, including teacher attrition.

4. MLOps & Productionization – Microsoft Fabric / Azure

• Develop and deploy ML solutions using Microsoft Fabric Notebooks, Python/R, Fabric Data Factory, OneLake Delta tables, and MLflow.

• Implement experiment tracking, model registry, versioning, and rollback mechanisms.

• Build reliable weekly batch-scoring pipelines for all active enrollments.

• Deliver ranked at-risk enrollment data to business users through Power BI dashboards using Direct Lake.

• Implement Dev/Test/Prod environments, Fabric deployment pipelines, and Azure DevOps CI/CD with approval gates.

• Manage secrets and dependencies using Azure Key Vault and pinned package environments.

• Establish monitoring for model drift, missing predictions, and data-quality failures, with alerts through Azure Monitor or Power Automate.

5. Measurement & Experimentation

• Design holdout and matched-control approaches to measure the effectiveness of retention outreach.

• Help distinguish genuine churn reduction from model accuracy alone.

• Contribute to experimentation initiatives involving matched control centers, difference-in-differences, effect sizes, and confidence intervals.

6. Technical Leadership

• Establish standards for reproducibility, code reviews, model reviews, documentation, and explainability.

• Mentor data scientists and engineers while collaborating with distributed client and delivery teams.

• Extend learnings from churn modeling to future use cases, including teacher attrition, occupancy and enrollment forecasting, and lead-to-enrollment scoring.

REQUIRED QUALIFICATIONS:-

• 7+ years of experience in data science or applied machine learning, including ownership of end-to-end ML outcomes in a senior or lead role.

• Proven hands-on experience in churn and retention analytics, including building and deploying models tied to measurable business results.

• Strong Azure data and ML experience is mandatory, with production ML delivery using Azure Machine Learning and/or Microsoft Fabric, Azure Data Factory/Fabric Data Factory, and MLflow-style model lifecycle management.

• Advanced Python skills, including pandas, scikit-learn, XGBoost/LightGBM, and libraries such as lifelines or scikit-survival.

• Advanced SQL skills and working proficiency in R.

• Strong statistical foundations in survival analysis, model calibration, hypothesis testing, experimental design, and causal inference fundamentals.

• Practical MLOps experience covering Git, CI/CD for ML, batch-scoring pipelines, model monitoring, and drift detection.

• Experience delivering model outputs to business users through Power BI.

• Excellent communication skills and direct experience working with US-based clients and presenting to senior business stakeholders.

• Ability to collaborate effectively across distributed, cross-functional teams with regular US Central Time overlap.

NICE TO HAVE:-

• Hands-on Microsoft Fabric experience with OneLake, Lakehouse/Delta, Fabric Notebooks, deployment pipelines, and Direct Lake.

• Domain experience in childcare, education, healthcare, subscription businesses, or multi-location consumer services.

• Experience with workforce attrition modeling, time-series forecasting (Prophet/ARIMA), or lead scoring.

• Familiarity with quasi-experimental methods, including Dynamic Time Warping matching, difference-in-differences, and synthetic controls.

• Experience handling sensitive or child-related personal data under privacy regulations such as COPPA.

• Exposure to Databricks, AWS SageMaker, GCP Vertex AI, or Azure ML online endpoints.

• Consulting or professional services delivery experience.

• Relevant certifications, such as Microsoft Fabric Analytics Engineer (DP-600), Azure Data Scientist Associate (DP-100), or Azure AI Engineer.

WHY JOIN NEUROGENT.AI

• Own a high-impact enterprise ML use case: Lead a flagship churn prediction program with a clear, measurable business objective.

• Build from the ground up: Help establish a scalable data science and ML foundation on Microsoft Fabric.

• Influence business outcomes: Turn predictive insights into practical interventions that help improve family retention.

• Shape the roadmap: Extend your work into future applications such as teacher attrition, enrollment forecasting, and lead scoring.

• Take real ownership: Work in an AI-first environment with direct client exposure, technical leadership, and opportunities to influence product and delivery decisions.

Interested in building ML solutions that deliver measurable business impact

Apply to join Neurogent.ai and help transform enterprise data into actionable decisions.