Sai Kumar Dwivedi

Sai Kumar Dwivedi is a PhD candidate and Doctoral Researcher at the Max Planck Institute for Intelligent Systems (MPI-IS) in Tübingen, Germany, working under the supervision of Dr. Michael Black and Dr. Dimitrios Tzionas in the Perceiving Systems department. He closely collaborates with Dr. Cordelia Schmid and has established himself as a leading researcher in computer vision, with multiple first-author publications in top-tier conferences.

His educational journey includes:

• PhD at International Max Planck Research School for Intelligent Systems (2021-Present)
• Master's and Bachelor's in Computer Science and Engineering from National Institute of Technology, Rourkela (2011-2016)

Dwivedi's research focuses on human perception in in-the-wild scenarios and human-object interactions. His work bridges 2D image understanding and 3D scene reconstruction, with particular emphasis on estimating human pose and shape from monocular images. He has pioneered approaches that leverage foundational models to address challenges in human-object contact estimation, creating methods that significantly advance the state-of-the-art in 3D human-object interaction understanding.

His publication trajectory shows a clear evolution from basic human pose estimation to sophisticated reasoning about human-object interactions using vision-language models. Recent work demonstrates how large foundational models can be adapted to understand and reconstruct 3D interactions from single RGB images, with applications spanning autonomous vehicles, robotics, and augmented reality systems.

Notable scientific achievements include:

• Outstanding Reviewers Award for CVPR 2025
• InteractVLM winning the Human Contact Challenge at CVPR 2025
• Outstanding Reviewers Award for ECCV 2024
• IMPRS-IS Scholar (2022)
• RA Fellowship (2021)

While not currently supervising PhD students, Dwivedi has mentored junior researchers through his work at MPI-IS and previously at Mercedes-Benz where his research was deployed in production vehicles (EQS 2021 and S-Class 2021). His research has been supported by the International Max Planck Research School for Intelligent Systems and the Max Planck ETH Center for Learning Systems.

As a key contributor to the Perceiving Systems department, Dwivedi's work on human-object interaction has practical applications in automotive safety systems, demonstrating the real-world impact of his research. His GitHub repository for InteractVLM has garnered significant attention with 111 stars, reflecting the community's interest in his innovative approaches to 3D interaction reasoning.