Machine Learning Engineer

This is a Full Time Job

#WeAreParamount on a mission to unleash the power of content… you in?
We’ve got the brands, we’ve got the stars, we’ve got the power to achieve our mission to entertain the planet – now all we’re missing is… YOU! Becoming a part of Paramount means joining a team of passionate people who not only recognize the power of content but also enjoy a touch of fun and uniqueness. Together, we co-create moments that matter – both for our audiences and our employees – and aim to leave a positive mark on culture.

Overview

As a member of the Applied Machine Learning Group, you’ll help build a world-class streaming experience within the team. Your mission is to get the user watching. You own the ''onboarding'' and ''re-entry'' experience, including high-commitment surfaces like Your Next Watch (YNW), Jump Back In (JBI), and While You Were Away (WYWA).

You’ll put customers first by using MLFlow for experiment tracking, Qdrant for retrieval, and Post-training RL to ensure that the content we surface at the start of a session is what the user is most likely to commit to. You will be in charge of implementing the components that solve the ''Cold Start'' problem for new users and provide seamless re-entry for returning ones. You’ll work in an IC2 (Mid-level) role to ensure these surfaces, which optimize for Start Rate under uncertainty, identify if a user stays or leaves within the first 30 seconds of app launch.

Responsibilities
• Independent Delivery: Own the implementation of features for JBI (Jump Back In) and YNW (Your Next Watch).
• Retrieval Optimization: Use Qdrant to find high-relevance candidates for re-entry carousels based on session history and global trends.
• Experiment Lifecycle: Use MLFlow to manage, track, and deploy experiments, ensuring a high bar for reproducibility.
• Model Training: Develop and serve models in GCP using TensorFlow/PyTorch, incorporating Post-training RL for reward-based optimization.
• Collaborative Quality: Participate in design reviews to ensure your components share the same high-reliability standards as the rest of the pod.
• Solving Cold Start: Build the logic that makes the app feel personalized even for users we know very little about.
• High-Commitment Accuracy: Optimize for the ''Play'' button–the ultimate signal of user commitment.

Basic Qualifications
• 3 years in MLE.
• Experience with GCP and MLFlow.
• Proficiency in TensorFlow/PyTorch.
• Knowledge with Vector DBs.
• Proven ability to implement features that solve the ''Cold Start'' problem for new users.
• Experience managing, tracking, and deploying experiments to ensure a high bar for reproducibility.

Bonus Skills
• Background in multi-stage ranking or ''Cold Start'' problems.
• Experience with retrieval optimization and re-entry carousels.
• Knowledge with Post-training RL for reward-based optimization.

#LI-JJ1

ADDITIONAL INFORMATION