Mengkang Hu is the founder of Evolvent AI (evolvent.co), where he is building infrastructure for recursive self-improvement. During his doctoral studies, his research focused on large language model–based agents and embodied AI. He authored or co-authored over 40 papers during this period, including more than 10 AI papers as first or corresponding author at leading conferences such as ICLR, KDD, and NeurIPS. His work has received over 3,000 citations on Google Scholar. He has been selected for the Tencent Qingyun Program, Xiaohongshu’s Top Talent Program, Microsoft Research Asia’s Stars of Tomorrow, and Sequoia China AI25. He led the team behind OWL, an open-source project with over 20,000 GitHub stars and one of the fastest-growing projects by star count in the AI agent space. Projects he has contributed to have collectively earned over 30,000 GitHub stars.

Evolvent AI is dedicated to building AI Scientist systems capable of continuous learning and self-evolution. Its current focus is on developing data and environments for LLMs and agents, and advancing practical applications of self-evolving agents in AI for Science (AI4Science), AI for AI (AI4AI), and related fields.

胡夢康博士期間從事large language model-based agent和embodied ai方向研究，現從事recursive self improving infrastructure方向創業 Evolvent AI 創始人。（evolvent.co）。就讀期間產出40+篇論文，其中10+篇一作/通訊AI論文（ICLR, KDD, NeurIPS等）。谷歌學術引用量3000+。曾入選騰訊青雲計劃，小紅書頂尖人才計劃，微軟亞洲研究院明日之星，紅杉中國AI25等。帶領團隊開發開源項目OWL，Github星標累計2w+，是Agent方向星標增速最快的項目之一。個人參與項目總計獲超過3w星標。

Evolvent AI 致力於打造能夠持續學習、自我進化的 AI Scientist 系統，現階段聚焦於面向 LLM 與 Agent 的數據和環境建設，並推動自進化 Agent 在 AI4Science、AI4AI 等領域的實際落地。