Senior Applied Scientist
About Crunchyroll
Founded by fans, Crunchyroll delivers the art and culture of anime to a passionate community. We super-serve over 100 million anime and manga fans across 200+ countries and territories, and help them connect with the stories and characters they crave. Whether that experience is online or in-person, streaming video, theatrical, games, merchandise, events and more, it’s powered by the anime content we all love. Join our team, and help us shape the future of anime!
Role Summary
We are hiring an Applied Scientist to help advance personalization across the Crunchyroll ecosystem. In this role, you will lead the scientific development of recommendation, ranking, and decisioning solutions that improve how fans discover and engage with anime series/movies, manga, merchandise, games, and other areas in the anime fandom.
Core Areas of Responsibility
Lead the research and development of recommendation, ranking, retrieval, and personalization methods tailored to Crunchyroll use cases across streaming, manga, ecommerce, and lifecycle marketing surfaces.
Frame ambiguous business and product questions into clear scientific problems, hypotheses, success metrics, and experimentation plans.
Design and run robust offline evaluation frameworks for recommender systems, including relevance, diversity, novelty, coverage, calibration, and long-term value metrics.
Partner with Product, Analytics, and Engineering to define online experiments, interpret results, and turn learnings into roadmap decisions and model improvements.
Develop user, content, and contextual understanding through feature design, representation learning, segmentation, and behavioral analysis.
Prototype and evaluate a range of approaches, including collaborative filtering, content-based methods, sequence modeling, deep learning, bandits, causal or uplift methods, and LLM-enabled recommendation techniques where appropriate.
Analyze user feedback loops and cross-domain interactions to improve discovery across video, merchandise, manga, and other ecosystem experiences.
Work closely with Machine Learning Engineers to translate promising research into production-ready solutions on our in-house recommendation platform.
Communicate scientific findings, model tradeoffs, and business implications clearly to technical and non-technical stakeholders.
Help establish best practices for experimentation, reproducibility, model governance, and scientific documentation within the personalization and recommendations function.
Qualifications
Experience: 5+ years of experience in applied machine learning, recommendation systems, search/ranking, experimentation, or a closely related area, with a track record of driving measurable product impact.
Scientific Depth: Strong foundations in machine learning, statistics, experimental design, and causal thinking, and the ability to choose the right level of modeling complexity for the problem at hand.
Recommendation Expertise: Hands-on experience with collaborative filtering, retrieval and ranking systems, representation learning, sequence / generative models, bandits, graph methods, or personalization for consumer products.
Technical Skills: Proficiency in Python and comfort with common ML libraries such as PyTorch, TensorFlow, Scikit-learn, XGBoost, or similar tooling. Experience working with SQL, distributed data processing, and cloud-based ML workflows is strongly preferred.
Experimentation Mindset: Knowledge of designing offline and online evaluations, reasoning carefully about metrics, and connecting experimental findings to user and business outcomes.
Cross-Functional Collaboration: Experience partnering effectively with engineers, product managers, analysts, marketers, and business stakeholders to move from idea to execution.
Communication Skills: Ability to explain sophisticated modeling decisions and ambiguous findings in a clear, decision-oriented way to diverse audiences.
Education: MS or PhD in Computer Science, Machine Learning, Statistics, Operations Research, Economics, or a related quantitative discipline, or equivalent applied industry experience.
Nice to Have
Experience personalizing content, commerce, media, entertainment, gaming, or subscription products at scale.
Familiarity with recommender-system failure modes such as popularity bias, cold start, sparse feedback, and feedback loop effects.
Experience with multi-objective optimization, constrained ranking, or balancing short-term engagement with long-term user value.
Exposure to generative AI, representation learning, or LLM applications that support recommendation and personalization workflows.
Published research, patents, or open-source contributions in recommendation systems, personalization, applied machine learning, or experimentation.