AllSpark is Xiaohongshu's AI research team advancing the frontier of agent intelligence through unified, general-purpose multimodal Agent Foundation Models.
We explore how agents learn, reason, and improve through mid-training, supervised fine-tuning, reinforcement learning, and on-policy distillation.
We bring these capabilities to products across Xiaohongshu, helping users discover, create, and connect every day.
We recruit Research Interns year-round. If you are interested in joining AllSpark, please send your CV to allspark@xiaohongshu.com or apply through the Xiaohongshu recruitment portal.
𝕏 X.com · 🤗 Hugging Face · 🧠 Models · 📄 Papers · 🔴 RedNote (小红书)