Principal Machine Learning Engineer

Paramount•New York, NY

38d•$233,600 - $350,400

About The Position

We are looking for a Principal Machine Learning Engineer to set the technical direction for personalization and discovery across our global streaming platforms. Your mission is to define the long-term architecture and modeling strategy. This strategy will help millions of viewers find films, series, live sports, and news. It will include Paramount+, Pluto TV, and future streaming products. You will focus on recommendations, ranking, retrieval, and real-time user awareness. This is a Principal role, meaning you are a senior-most technical authority in the org. You set multi-quarter technical strategy, drive cross-team alignment, and are accountable for the scientific rigor of how we model long-term user satisfaction. You will work in a GCP-based environment. You will use TensorFlow and PyTorch. The focus will be on modern personalization techniques. These techniques include representation learning, multi-modal embeddings, contextual bandits, and session modeling.

Requirements

8+ years of experience in MLE, applied science, or large-scale recommender/ranking systems, with a track record of setting technical direction.
Proficient in representation learning. Experience with multi-modal embeddings. Knowledge of contextual bandits and session modeling.
Robust fluency in experimentation methodology, A/B testing, causal reasoning, and metric design.
Proficiency in GCP, TensorFlow, and PyTorch.
Demonstrated ability to influence technical strategy across multiple teams.

Nice To Haves

Experience in high-traffic, real-time streaming or consumer apps.
Published work or recognized contributions in ranking, recommender systems, or applied ML.

Responsibilities

Set Technical Strategy: Own the multi-quarter technical roadmap for personalization, covering candidate generation, ranking, and exploration.
Architect end-to-end systems. Design multi-stage personalization systems. These systems include retrieval, deep ranking, contextual embeddings, and bandit-based exploration. You will manage the entire process, from feature engineering to training, serving, and monitoring.
Advance modeling in production. Drive advanced techniques. Use representation learning, multi-task learning, multi-modal embeddings, and session modeling. These strategies will help improve user satisfaction.
Cross-Pod Influence: Partner with Core Science, Content Understanding, ML Platform, and Product to align personalization with broader strategy.
Operate at Scale: Ensure personalization pipelines are high-throughput, reliable, and observable in GCP using TensorFlow/PyTorch and big-data tooling (Beam, BigQuery).
Raise the quality of experimentation. Establish practices that ensure high integrity in experiments. Improve the correlation between offline and online results. Guide feature rollouts with solid scientific methods.
Mentorship & Talent: Mentor engineers, set technical standards across the org, and grow the next generation of senior ML talent.
Mitigate Systemic Risk: Identify and address feedback loops, exposure biases, and filter-bubble dynamics in how content is surfaced.