Head of Data Science — AI-Driven Behavioral Embeddings

Jobtome; West New York (NJ); On-site; USD 263,000 - 394,000; Full time

Job summary

Spectrum is seeking a Head of Data Science to lead the development of transformer-based behavioral embeddings and predictive models that power audience, insights, and measurement products. You will own end-to-end data science programs, establish responsible AI standards, and steer collaboration with product, sales, and partners to drive pilots to contracts.

You will manage a team of data scientists and ML engineers, define hiring priorities, and ensure production deployment with a bias toward

Qualifications

• Proven record leading data science for data, audience, identity, or measurement products.
• Hands-on predictive modeling on consumer data at scale.
• Strong grounding in measurement and causal inference.
• Experience with privacy-preserving analytics and data regulation.
• Ability to explain methodology to executives and buyers.

Responsibilities

• Build predictive models behind the platform's audience and data products.
• Own the platform's Feature Store as a commercial product.
• Lead the evaluation, MLOps, and responsible AI standards.
• Direct the roadmap for the embedding model with external partners.
• Build and lead a data science team aligned to revenue impact.

Skills

Leadership

Predictive modeling

Causal inference

Privacy by design

Python & SQL

Transformers embeddings

Cloud platforms

Team building

Education

Bachelor's in Quant

Tools

Snowflake

LiveRamp

AWS

Spark

Databricks

UID2/RampID

Job description

Spectrum is seeking a Head of Data Science to lead the development of transformer-based behavioral embeddings and predictive models that power audience, insights, and measurement products. You will own end-to-end data science programs, establish responsible AI standards, and steer collaboration with product, sales, and partners to drive pilots to contracts.

You will manage a team of data scientists and ML engineers, define hiring priorities, and ensure production deployment with a bias toward