Principal Core Science Machine Learning Scientist

Overview

We are seeking a Principal Machine Learning Scientist to join the Core Science Group. This is a foundational leadership role designed for a visionary who wants to build the "atomic units" of our next-generation personalization ecosystem. Your mission is to define and develop Personalization Foundation Models and the ML Primitives that will power every surface across the Paramount streaming portfolio.

As a Principal Scientist, you operate at an organizational scale, solving the most complex and ambiguous challenges in model architecture and systems engineering. You will lead the design of agentic components for personalization products and push the boundaries of how we think about the "AI Kernel." This role requires a rare blend of deep scientific research and the low-level systems engineering (Rust, custom kernels) needed to make foundation models performant at a global scale.

• Defining the Foundation: You create the core models that every Applied ML pod will use as their starting point, ensuring we don't reinvent the wheel for every surface.
• Architecting Agency: You lead the transition from passive recommendation to autonomous agentic components that can reason about content and user intent.
• Systems Innovation: By pushing into Helion kernels and Triton optimization, you define the efficiency and cost-profile of our entire ML infrastructure.

• Drive Long-term Strategy: Own the technical direction for Personalization Foundation Models, setting a 2–3 year horizon for research and implementation.
• Develop ML Primitives: Build the shared libraries and foundational components used by ranking and retrieval systems across the organization.
• High-Performance Engineering: Implement performance-critical components in Rust and optimize inference using Triton and custom Helion kernels.
• Agentic Research: Design the orchestration and reasoning logic for agentic personalization components.
• Influence & Mentorship: Shape the roadmap for both the Core Science and Applied ML groups, influencing senior leaders and mentoring the company's top technical talent.

• 10+ years of experience in ML research and systems engineering; mastery of PyTorch and Rust; deep experience with inference optimization (Triton).

• Published research in Foundation Models or Agentic Systems; experience building or contributing to AI Kernels;
  PhD in Computer Science or related field.

• 12 Months: Delivery of the first "Paramount Personalization Primitive" adopted by at least two major Applied ML pods, resulting in measurable improvements in training efficiency or model accuracy.