SiriusXM is looking to protect its platform from streaming fraud, unfair royalty payouts, and degraded user experiences by researching and developing ML models to detect anomalous user behavior and build foundational ML/AI systems that understand content.
Requirements
- Deep understanding of supervised and unsupervised machine learning, anomaly detection, and graph-based techniques
- Hands-on experience with LLMs, including retrieval-augmented generation (RAG) pipelines, vector stores, and orchestration frameworks
- Experience with large-scale data and production ML pipelines (e.g., TensorFlow/PyTorch, Airflow, Spark, cloud platforms like AWS/GCP)
- 3+ additional years of research and development experience in real-world content understanding and/or fraud detection systems.
Responsibilities
- Research, design, and build high-impact machine learning models to combat streaming fraud at scale and extend SiriusXM’s content understanding capabilities.
- Develop content understanding models using ML/AI techniques that power downstream systems across listener experiences, royalty processing, and catalog management.
- Create new signals and pipelines to better model user behavior, content consumption patterns, and behavioral anomalies.
- Frame business goals as machine learning problems across Pandora and SiriusXM and own their execution.
- Collaborate with cross-functional partners including Engineering, Product, Content Licensing, and Curation
Other
- A Ph.D. or M.S. in a quantitative field and 3-5 years of relevant work experience
- Excellent written and verbal communication skills, with the ability to effectively advocate technical solutions to scientists, engineers, and product audiences.
- Passion for data-driven research, development, and experimentation.
- Self-motivated, growth-oriented, and driven to pursue solutions to challenging problems.
- Must have legal right to work in the U.S.
