Machine Learning Research Intern (Summer 2027)

At Scale , our mission is to develop reliable AI systems for the world's most important decisions. Our products provide the high-quality data and full-stack technologies that power the world's leading models, and help enterprises and governments build, deploy, and oversee AI applications that deliver real impact. We work closely with industry leaders like Meta, Ernst & Young, Mayo Clinic, Time Inc., the Government of Qatar, and U.S. government agencies including the Army and Air Force. Scale's Machine Learning Research team builds the foundation for the next generation of AI systems — pushing the boundaries of what's possible with frontier models while ensuring safety, reliability, and alignment at every step. Our work spans generative AI, reinforcement learning and advanced post-training methods, scalable oversight, synthetic data pipelines, red teaming, and evaluation science. We are developing a large-scale hybrid human-machine system to power machine learning pipelines for dozens of industry-leading customers. These models and systems form the backbone of Scale's long-term strategy, enabling billions of tasks monthly and supporting some of the most complex and advanced use cases in the AI ecosystem. As an intern, you'll work on a combination of deeply technical ML applications in production and cutting-edge research problems, with mentorship from world-class researchers and the opportunity to collaborate with leading teams across industry and academia. Interns own real, published or shipped work — not side projects. Example Projects Measure the dangerous capabilities of frontier models and conduct preparedness research Research the science and creation of new benchmarks for frontier models Develop new methods for training models to excel on extremely difficult reasoning problems that require long chains of thought Research reinforcement learning and post-training methods that improve model reliability on high-stakes tasks Research scalable oversight protocols that enable humans to produce and quality control reasoning chains beyond their native capabilities Study the boundaries of model generalization and capabilities to inform data-driven advancements Research synthetic and hybrid data with humans in the loop to scale up high-quality data generation Take models currently in production, identify areas for improvement, improve them using retraining and hyperparameter searches, then deploy without regressing on core model characteristics Create post-training or agentic solutions that integrate into our ability to deliver applications for enterprise and public sector clients Required to Have Currently enrolled in a BS, MS or PhD program with a focus on Machine Learning, Deep Learning, or Natural Language Processing, with a graduation date in Fall 2027 or Spring 2028 Available for a Summer 2027 internship (May/June start dates) in San Francisco Or New York. Prior research experience or a track record of publications on LLMs, NLP, multimodal models, agents, safety, evaluation, alignment, or a related field Experience with one or more general purpose programming languages, including Python, Javascript, or similar Ideally You'd Have A previous internship in Machine Learning, Deep Learning, Natural Language Processing, Adversarial Robustness, Alignment, Evaluation, or Agents Experience as a researcher — through internships, full-time roles, or at a lab Publications at top-tier venues such as NeurIPS, ICML, ICLR, CVPR, ACL, EMNLP or AAAI, or contributions to open source projects Hands-on experience with reinforcement learning, agentic systems, or evaluation infrastructure How We Work Our credos shape how we make decisions and work as a team: Earn Customer Love — trust is earned with every interaction, every delivery, and every problem we solve Team Flow — ideas, energy, and support move freely across teams; optimize for the whole Quality is Our Cheat Code — t