Research Scientist • Intern

Who you are

You are pursuing a PhD and have a strong interest in evaluating large language models (LLMs) — your academic background has equipped you with the skills necessary to assess and improve AI systems. You are passionate about artificial intelligence and are eager to contribute to the development of robust evaluation frameworks that ensure the safety and reliability of cutting-edge technologies. Your experience in data analysis allows you to interpret complex datasets and draw meaningful conclusions that can influence the future of AI.

You have a solid understanding of benchmarking methodologies and are familiar with prompt engineering — you know how to design effective prompts that can test the capabilities of LLMs across various domains. Your analytical mindset enables you to identify areas for improvement in AI models, and you are committed to advancing the field through rigorous evaluation practices. You thrive in collaborative environments and are excited about the opportunity to work with a team of experts in the Meta Superintelligence Lab.

Desirable

Experience with machine learning frameworks and tools is a plus, as is familiarity with programming languages commonly used in AI research. You are open to learning new methodologies and technologies that can enhance your contributions to the team. A background in AI safety and reliability would also be beneficial, as it aligns with the goals of the lab.

What you'll do

As an LLM Evaluation Scientist Intern at Meta, you will engage in the evaluation and benchmarking of large language models, focusing on their performance across language and multimodal domains. You will collaborate with a team of researchers to develop methodologies that assess the capabilities of these models, ensuring they meet the highest standards of performance and safety. Your role will involve designing experiments that test various aspects of LLMs, analyzing the results, and providing insights that can guide future improvements.

You will have the opportunity to contribute to the development of tools that facilitate the evaluation process, making it easier to benchmark models against established standards. Your work will directly impact the advancement of AI technologies, as you help shape the methodologies used to evaluate their effectiveness. You will also participate in discussions about the implications of your findings, contributing to the broader conversation about AI safety and reliability.

What we offer

Meta offers a dynamic internship experience that allows you to work on meaningful projects that have a real impact on the future of AI. You will be part of a collaborative team that values innovation and encourages you to share your ideas. The internship duration ranges from twelve to twenty-four weeks, providing flexibility in your schedule. You will gain hands-on experience in a fast-growing organization and have the chance to learn from industry leaders in AI research. This is an exciting opportunity to kickstart your career in a field that is shaping the future of technology.