Data Scientist - Machine Learning

EXL; Gurugram District; On-site; INR 900,000 - 1,300,000; Full time

Job summary

EXL is seeking an experienced ML professional to design, train, and validate both supervised and unsupervised models using Python. You will own production deployment via REST APIs and cloud pipelines, and build robust ML workflows with PySpark, Airflow, MLflow, or Databricks.

You will implement drift detection, monitoring, and version control with Git/DVC, packaging with Docker, and collaborate with product and engineering teams to translate business problems into measurable outcomes.

Responsibilities

• Model Development: Design, train, and validate supervised and unsupervised ML models (e.g., regression, classification, clustering, time-series forecasting, NLP, or deep learning) using Python.
• Full-Stack ML / Deployment: Own model deployment into production environments via REST APIs, microservices, or cloud-native pipelines (AWS, Azure, or GCP).
• MLOps & Pipelines: Build and maintain reproducible data and feature pipelines using tools like PySpark, Airflow, MLflow, or Databricks.
• Containerization & Version Control: Version control code and model artifacts (Git, DVC, or MLflow) and package applications using Docker.
• Monitoring & Optimization: Implement model drift detection, logging, and performance monitoring to ensure high availability and accuracy in production.
• Stakeholder Collaboration: Translate complex business problems into quantitative metrics and present actionable insights to cross-functional product and engineering teams.

Job description

Key Responsibilities

• Model Development: Design, train, and validate supervised and unsupervised machine learning models (e.g., regression, classification, clustering, time-series forecasting, NLP, or deep learning) using Python.


• Full-Stack ML / Deployment: Own model deployment into production environments via REST APIs, microservices, or cloud-native pipelines (AWS, Azure, or GCP).


• MLOps & Pipelines: Build and maintain reproducible data and feature pipelines using tools like PySpark, Airflow, MLflow, or Databricks.


• Containerization & Version Control: Version control code and model artifacts (Git, DVC, or MLflow) and package applications using Docker.


• Monitoring & Optimization: Implement model drift detection, logging, and performance monitoring to ensure high availability and accuracy in production.


• Stakeholder Collaboration: Translate complex business problems into quantitative metrics and present actionable insights to cross-functional product and engineering teams.