Data Scientist with AWS Sagemaker Exp

Dice is the leading career destination for tech experts at every stage of their careers. Our client, Rocket, is seeking the following. Apply via Dice today!

Data Scientists

• Focused on replication, scaling, and supporting ML solutions across clients.

• Responsibilities:

• Deploy and operationalize models created by economists.

• Onboard new partners (80+ airline clients).

• Expand and refine existing solutions (e.g., reward improvements).

• Requirements:

• 4+ years of experience, degree in related field.

• Ability to become self-sufficient over time and eventually take over economist responsibilities.

Technical Stack & Tools

• Primary ML algorithms:

• Contextual Bandits (reinforcement learning)

• XGBoost (baseline predictive models)

• K-means clustering

• Linear regression

• Platforms:

• Amazon SageMaker (Studio)

• Redshift

• Snowflake

• Q for Business

• Languages:

• SQL

• Python

Use Cases & Deployment Strategy

• Primary use cases:

• Reinforcement learning for upgrade bidding.

• Example: Customer receives an upgrade offer suggest an alternative product/ancillary offer.

• Hospitality use case: Ancillary services offered at random models can optimize targeting.

• New partner onboarding:

• Begin with existing data (80 airline partners).

• Economist monitors and customizes model.

• Roll out partner by partner using shared framework.

Additional Notes

• Solutions must work in both B2B and B2C contexts.

• Human end-customers are always the recipient of offers, even in B2B partnerships.

• By 2025, all models will be built and supported using SageMaker libraries.

Must Haves :

Amazon SageMaker Studio _for ML development and orchestration.__Algorithms like:

_

• Contextual Bandits used for real-time decision making (e.g., pricing, recommendations).
• XGBoost a high-performance gradient boosting algorithm, often used in tabular data for predictions.
• K-means Clustering for unsupervised segmentation or grouping tasks.
• Linear Regression for basic predictive modeling or as a baseline.