Machine Learning Engineer Intern - (LLM Risk Surface) - 2025 Summer (PhD)

Tiktok - San Jose, CA

Full-time - Intern

San Jose, CA

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services

About the position

TikTok is the leading destination for short-form mobile video, and our mission is to inspire creativity and bring joy. The Business Risk Integrated Control (BRIC) team is dedicated to protecting TikTok users, including content consumers, creators, and advertisers, while securing platform health and ensuring community experience authenticity. This team is responsible for building infrastructures, platforms, and technologies, collaborating with various cross-functional teams and stakeholders to minimize the impact of inauthentic behaviors on TikTok's platforms, which include TikTok, CapCut, Resso, and Lark. The BRIC team addresses multiple community and business risk areas such as account integrity, engagement authenticity, anti-spam measures, API abuse, growth fraud, live streaming security, and financial safety related to ads and e-commerce. As an intern in this team, you will have a unique opportunity to gain first-hand exposure to the company's strategy in key security initiatives. You will be involved in developing scalable, robust, intelligent, and privacy-safe systems and solutions that are product-friendly. The challenges you will face are not typical day-to-day technical puzzles; instead, you will be part of a team that is creating novel solutions to unprecedented challenges within a rapidly evolving product ecosystem. The work environment is dynamic, requiring fast-paced and transferable solutions that make a significant impact. Internships at TikTok are designed to provide students with industry exposure and hands-on experience, allowing you to turn your ambitions into reality. You will have the chance to participate in a vibrant blend of social events and enriching development workshops, utilizing your knowledge in real-world scenarios while laying a strong foundation for personal and professional growth. This Internship Program runs for 12 weeks, beginning in May/June 2025, with specific start dates available. Successful candidates must clearly state their availability in their resumes and can apply to a maximum of two positions, with applications reviewed on a rolling basis. Candidates who pass the resume evaluation will be invited to participate in TikTok's technical online assessment in HackerRank.

Responsibilities

• Responsible for the research and development of algorithms for AI native applications and optimization of LLM (Large Language Models), driving the exploration and application of cutting-edge technologies.
• Enhance the capabilities of natural language understanding based on LLM, such as content understanding/classification, vector retrieval, structuring, and representation learning of short/long texts.
• Combine the most advanced LLM technology to summarize, understand, and classify user behavior/features.
• Explore the application of Agent in complex tasks and achieve the application of complex tasks in the risk control field based on LLM.

Requirements

• Currently pursuing a PhD in Computer Science, Computer Engineering, Math, Statistics or a related technical discipline.
• Able to commit to working for 12 weeks during Summer 2025.
• Good coding standards, a passion for writing code, and the ability to produce high-quality designs and code.
• Excellent problem analysis and problem-solving abilities, capable of identifying the essence of issues from complex data.
• Experience in algorithmic research and development of mainstream risk control strategies and LLM systems development experience is a plus.
• Strong sense of responsibility, proactive attitude, and good communication and teamwork skills.

Nice-to-haves

• Graduating December 2025 onwards with the intent to return to degree program after the completion of the internship.
• Priority will be given to those who master the algorithm principles of LLM (Large Language Models), Fine-tuning, Prompt Engineering, vector databases, and Agent application frameworks such as LangChain/Autogen; familiarity with the latest LLM technologies and experience in training or applying them is a plus.

Benefits

• Hands-on experience in developing fundamental skills and exploring potential career paths.
• Vibrant blend of social events and enriching development workshops.