Lead Machine Learning Engineer, Entry and Re-engagement

Paramount•New York, NY

About The Position

We are looking for a Senior Lead Machine Learning Engineer for the Entry pod. This team owns the high stakes surfaces that establish whether a user stays or bounces within seconds of opening the app. You will lead the ML strategy for Onboarding, Re-entry, and High-Commitment Recommendations, including core features like "Your Next Watch", "Jump Back In", and "While You Were Away". The technical task of this pod is optimizing start rates under high uncertainty. You will design systems that oversee the "cold start" problem for new users and the "intent gap" for returning users. Because these surfaces share similar failure modes such as over-indexing on recency or failing to surface "must-watch" content. You will architect an unified approach to re-engagement that balances historical preference with real-time context. The Entry pod owns the "Moment of Truth." In this role, you will directly shape: The First Touchpoint: Building the onboarding models that turn a first-time visitor into a long-term subscriber. The Re-entry Loop: Perfecting the "Jump Back In" experience to ensure users can resume their journey with zero friction. High-Stakes Discovery: Owning "Your Next Watch" (YNW), the primary engine for transitioning a user from a finished series into their next obsession.

Requirements

6-8+ years of experience in machine learning engineering, specifically in ranking, retrieval, or reinforcement learning.
Proven experience building recommendation systems that perform under data sparsity or for "new-to-system" entities.
Deep knowledge of multi-stage ranking, learning-to-rank (LTR), and handling temporal features in real-time.
Deep expertise in designing and deploying scalable hybrid retrieval architectures capable of processing massive interaction volumes in real-time.
Proficiency in A/B testing and the design of complex online metric frameworks, including primary, secondary, and guardrail indicators to validate model impact.
Proficiency in leveraging modern processing frameworks like Spark, Beam, or BigQuery to engineer features and productionize machine learning models at massive scale.
Demonstrated ability to drive technical initiatives collaboratively with Product, Engineering, and Data Science partners to deliver on shared business objectives.
Mastery of Python, Neural networks (Pytorch, tensorflow, JAX), Distributed systems (Ray or Spark), Orchestration (e.g., Airflow, Argo Workflows), MLOps (MLFlow/ Weights and Biases), DevOps (Docker, Kubernetes), Ranking, Search, Recommender systems, Java (nice to have), and Serving Frameworks (Nvidia Triton).
Experience mentoring junior engineers or leading technical initiatives, with a track record of spearheading complex ML projects from research to global production.

Nice To Haves

Experience with Sequential/Session-based models (RNNs, Transformers, Recommendation Systems, Search, Ranking) for predicting the "next" best action.
Background in Exploration/Exploitation (Bandits) for handling user uncertainty.
Knowledge with Causal Inference to distinguish between organic re-entry and model-driven lift.
Experience in high-scale consumer tech (Streaming, E-commerce, or social media).
Advanced proficiency in re-ranking methodologies focused on balancing content diversity with top-tier relevance.

Responsibilities

Contribute to the technical vision for re-engagement and onboarding, leading a pod of senior engineers to deliver high-impact production models.
Architect models for "Jump Back In" (JBI) and "While You Were Away" (WYWA) that account for temporal decay, episode progress, and cross-device signals.
Develop advanced onboarding algorithms that use minimal metadata and global trends to provide high-quality recommendations to new users immediately.
Develop multi-stage retrieval pipelines that successfully merge traditional feature-driven methods with semantic vector search, ensuring seamless integration with downstream ranking models for optimal performance.
Lead the development of the "Your Next Watch" (YNW) engine, optimizing for long-form commitment rather than just a click.
Implement exploration/exploitation strategies to navigate the uncertainty of user intent during app entry.
Define and optimize for Start Rate, Time to Play, and Day-1 retention. Beyond business KPIs, you will own the system-level Service Level Objectives (SLOs), ensuring high throughput and low-latency delivery of recommendations in production.