Wenwei Zhang,张文蔚

Wenwei Zhang is a Young Research Scientist at Shanghai Artificial Intelligence Laboratory. He works on post-training and self-improvement (scalable oversight) of foundation models, including Intern-S1/2, with a focus on AI Agents.

Current focus
Self-improving foundation models and agents via interaction and scalable oversight
Open positions
PhD/postdoc/intern and full-time researcher roles are open.
2023-

Current work at Shanghai AI Laboratory on post-training, agents, and scalable oversight.

2019-

Core member of OpenMMLab; built and released MMDetection3D and has led MMDetection/MMDetection3D development since 2020.

Ph.D.

MMLab, Nanyang Technological University, Singapore, supervised by Professor Chen Change (Cavan) Loy.

Research Direction

Self-improvement of Intelligence

My research focuses on foundation models and agents that improve through interaction with agents, environments, tools, and humans, under scalable oversight. I study two connected questions:

Machine Cognition

How can machines model, learn, and continually evolve cognitive capabilities in the human world?

Human-Machine Collaboration

How can evolving machine cognition assist complex human activities while keeping humans valuable in an increasingly automated world?

Recent News

Selected updates

  1. Intern-S1, state-of-the-art open-source multi-modal foundation model has been released. See Intern-S1 Collections and tech report.

  2. InternLM2.5-StepProver won the Best Paper Runner-Up Award in ICML2025 AI4Math Workshop.

  3. InternLM2.5-Chat, state-of-the-art open-source LLMs in opencompass and huggingface leadboard, have been released. See InternLM2.5 Model Zoo.

  4. InternLM2-Chat, state-of-the-art open-source LLMs in opencompass and AlpacaEval2, have been released. See InternLM2 Model Zoo.

  5. 2 papers (CLIPSelf and UniHSI) are accepted as spotlights by ICLR2024.

  6. We release OpenMMLab 2.0 with a new core, MMEngine.

  7. Video K-Net is accepted by CVPR 2022 (oral).

  8. K-Net is accepted by NeurIPS 2021. Code has been released at here.

  9. MMDet3D team obtains Best PKL Award and best vision-only results in the 3rd nuScenes detection challenge of 5th AI Driving Olympics, NeurIPS 2020. Paper of our multi-modality method is released in arxiv.

  10. Second runner up in LVIS2020 Challenge. Paper can be found in arxiv.

  11. We release MMDetection3D, OpenMMLab's next-generation platform for general 3D object detection.

  12. Win the 1st prize in COCO 2019 Object Detection Challenge (no external data). (Team: MMDet)

Academic Service

Community roles

  • Area Chair: CVPR2026.
  • Conference Reviewer: CVPR2020-2025, ICCV2021-2025, ECCV2020-2024, ICLR2022-2026, NeurIPS2021-2025, ICML2023-2024, ACM MM2020.
  • Committee member and speaker of OpenMMLab Tutorials in CVPR 2021/2022/2023, and AAAI2023.