AI Safety Policy Intern: Multimodal Risk and Evaluation

BYTEDANCE PTE. LTD.; Singapore; On-site; SGD 18,000 - 23,000; Full time

Job summary

ByteDance invites a Project Intern to contribute to the AI Data and Safety team, focusing on safety evaluation, policy research, and data-driven improvements across Seed models.

You will participate in research on AI safety developments, help design frameworks for multi-model safety assessments, and assist with data generation pipelines—exposure to real-world challenges in a fast-paced tech environment.

Qualifications

• Currently pursuing Bachelor's degree in AI Policy, Public Policy, Political Science, International Relations, Law, Philosophy, or a related discipline.
• Strong analytical skills to interpret qualitative and quantitative data and translate them into clear insights.
• Creative problem-solving mindset, with comfort working under ambiguity and leveraging tools and technology to improve processes and outputs.

Responsibilities

• Help to conduct research on the latest developments in AI safety across academia and industry, and support the identification of limitations in existing evaluation paradigms.
• Aid in designing and continuously refining safety evaluation frameworks for multi-models to assess safety-related behaviors, failure modes, and alignment with responsible AI principles.
• Support projects to enforce safety training or evaluate safety data, which may include data analysis to uncover insights that will inform model iteration and product design improvements.
• Support the designing, testing, and implementation of automated workflows and methodologies that ensure scalability, efficiency, and coverage within safety contexts, including for safety policy research, safety data generation, data analysis, and evaluation framework validation purposes.

Job description

ByteDance invites a Project Intern to contribute to the AI Data and Safety team, focusing on safety evaluation, policy research, and data-driven improvements across Seed models.

You will participate in research on AI safety developments, help design frameworks for multi-model safety assessments, and assist with data generation pipelines—exposure to real-world challenges in a fast-paced tech environment.