Research Scientist Graduate (Monetization Technology - Business Integrity) - 2027 Start (PhD)

Our Business Integrity team has a strong user focus and a dedication to technical excellence. We aim to meet our users’ needs with reliable and high-performing platforms and services. We are looking for strong machine learning engineers who are excited to grow their business understanding, build highly scalable machine learning models, and partner across disciplines with global teams, in pursuit of excellence. Given the fast growth of TikTok in the world, we are working on building a next-generation content understanding system for TikTok monetization. We are seeking Research Engineers who are experienced in machine learning, which can help us create an ecosystem that rewards high-quality user experience and advertiser value.

Topic Content: With the explosive growth of digital content, intelligent moderation has become a core capability for internet platforms. However, as moderation scenarios grow increasingly complex and adversarial tactics continue to evolve, traditional approaches are facing unprecedented challenges. The current landscape is characterized by multiple technical difficulties, including the dynamic nature of moderation rules, content complexity, sample scarcity, escalating adversarial behaviors, and a lack of interpretability. In particular, existing open-source large models often do not perform as effective as we expect in scenarios involving evolving moderation rules, long-form text, long temporal sequences, multi-languages, limited sample data, and adversarial content generated by AIGC.

We are looking for talented individuals to join our team. As a graduate, you will get opportunities to pursue bold ideas, tackle complex challenges, and unlock limitless growth.

Successful candidates must be able to commit to an onboarding date by the end of the year. Please state your availability and graduation date clearly in your resume.

Responsibilities

In this role, you will build a leading moderation system that enables end-to-end capabilities for accurate rejection decisions, interpretable reasoning, and intelligent remediation, achieving fully automated moderation with performance surpassing human benchmarks.

Qualifications

Minimum Qualifications

• Individuals who are completing or have recently completed a PhD degree in Computer Science, Software Engineering, Electronic Engineering, Automation, Mathematics, Statistics, or a related technical discipline (completed, or expecting to graduate within 12 months).
• Research experience in one or more of the following areas — Ads, Search, Recommender Systems, NLP, CV, Multimodal, or Agent technologies — demonstrated through publications, thesis work, internships, or research projects.
• Solid understanding of large language model methodology, including pre-training, post-training/alignment, and evaluation, with the ability to read, reproduce, and critique recent literature.
• Ability to independently formulate a research problem, design controlled experiments, and draw sound conclusions from ambiguous results.
• Proficient in Python and at least one deep learning framework such as PyTorch or TensorFlow.
• Strong problem-solving ability, a collaborative mindset, and a genuine interest in translating research into real-world products.

Preferred Qualifications

• Publications at top-tier venues (ICLR, NeurIPS, ICML, ACL, EMNLP, CVPR, ICCV, ECCV, TPAMI), or equivalent evidence of research ability.
• Experience with RLHF / post-training, reasoning, or novel model architecture design.
• Experience designing evaluation benchmarks or LLM/VLM-as-Judge methodologies.
• Experience building agentic systems, RAG pipelines, or multimodal applications.
• Track record of research that shipped into a production system.