About the role#
This internship focuses on building AI-native risk intelligence systems for TikTok. You will work on transitioning risk control from traditional rule-based models to multi-agent architectures. The team emphasizes tool-augmented reasoning, modular skill composition, and feedback-driven system evolution to address challenges in large-scale content production.
What you'll need#
- Current enrollment in a PhD program for Computer Science, Computer Engineering, or a related field.
- Proficiency in deep neural networks, ensemble trees, reinforcement learning, graph theory, transfer learning, and unsupervised learning.
- Experience with Python, SQL/Hive, and Hadoop.
- Strong grasp of algorithms and data structures.
- Analytical skills, attention to detail, and time management.
- Software engineering experience from past internships, publications, or coding competitions is preferred.
- Prior experience in e-commerce or risk control is a plus.
What you'll do#
- Develop and implement machine learning algorithms to manage business risks across TikTok products.
- Prototype novel solutions and run experiments to validate hypotheses.
- Build data querying infrastructure for both real-time and offline analysis.
- Collaborate with multidisciplinary teams to improve automation processes.
- Define risk control metrics and align stakeholders on numeric goals.
- Document technical processes and present results using data visualizations.
Location & details#
- Location: Seattle, Washington.
- Type: Full-time internship.
- Schedule: Rolling basis.
- Modality: On-site.


