Research Scientist Intern (Ads Integrity) - 2026 Start (PhD)

Responsibilities

• Lead research and development of advanced generative AI technologies, including LLMs, multimodal models (text/image/video), and deepfake detection/synthesis, focusing on optimizing performance across pre-training, SFT, RLHF, and AI safety.
• Design and deploy cutting-edge AIGC solutions for content understanding and monetization in diverse applications such as ads, e-commerce, short video, and live streaming, contributing to the creation of next-generation AI-driven ecosystems.
• Drive advancements in LLM-based agents using reinforcement learning to enable autonomous reasoning, planning, and interactive capabilities, addressing real-world challenges in dynamic environments.
• Innovate techniques to improve the efficiency of large-scale model training and inference, including distillation, quantization, and speculative decoding, for scalable and practical deployment in production.
• Collaborate with interdisciplinary teams to transition research breakthroughs into production-grade AI services, ensuring robust, low-latency, and cost-effective solutions.
• Stay at the forefront of generative AI research by contributing to patents, publications, and open-source projects, while actively monitoring and contributing to the latest industry trends and innovations.

Qualifications

• Current Ph.D. student in Computer Science, AI, Machine Learning, or related fields by 2026 (or equivalent industry experience).
• Strong foundational experience in deep learning, NLP, and generative models (LLMs, diffusion models, etc.).
• Hands-on experience with large-scale model training, RLHF (Reinforcement Learning from Human Feedback), and multimodal learning (text, image, video).
• Proficiency in one or more deep learning frameworks such as PyTorch, JAX, or TensorFlow, with familiarity in distributed training frameworks.

Preferred Qualifications

• A track record of publications or active research/paper review at top-tier conferences (NeurIPS, ICML, ACL, CVPR, etc.) or equivalent.
• Knowledge of AI safety, alignment, and adversarial robustness, with an interest in responsible AI development.
• Experience in developing agentic systems utilizing reinforcement learning.
• Strong engineering skills with the ability to deploy models at scale and optimize for performance.