James O'Brien

Timeline

1. 1992–2000

   Computer science degrees

   He earns a BS in computer science at Florida International University in 1992. At the Georgia Institute of Technology, he completes an MS in computer science in 1997 and a PhD in computer science in 2000.

2. 2000–2016

   Berkeley and SIGGRAPH roles

   He joins Berkeley's computer science department in 2000. By 2015, he leads the Berkeley Computer Animation Modeling Group. He serves as a director at large of ACM SIGGRAPH from 2010 to 2016.

3. 2013

   Detecting inconsistent shadows

   In 2013, he coauthors “Exposing Photo Manipulation with Inconsistent Shadows” with Eric Kee and Hany Farid. The method uses geometry and linear programming to test whether shadows in a photograph are physically consistent, helping detect photo manipulation.

4. 2015

   An Academy technical award

   In 2015, he receives a Technical Achievement Award from the Academy of Motion Picture Arts and Sciences for finite element research used in the Digital Molecular Matter toolkit. The toolkit simulates the motion of objects and structures as they are damaged or destroyed.

5. 2018

   Deep learning for character rigs

   With Stephen Bailey, Dave Otte and Paul DiLorenzo, he coauthors “Fast and Deep Deformation Approximations” in 2018. The method uses deep learning to approximate nonlinear character deformations and reduces deformation computation time by a factor of five to ten in its production rig examples.

6. 2021

   Testing the Picture-O-Meter

   In 2021, he coauthors a study with Cuihua Shen and Mona Kasra that tests the Picture-O-Meter image authenticity label with 2,440 participants. Images marked “Altered” receive lower credibility ratings than images with no label or an “Un-altered” label.

7. 2023

   Identifying virtual reality users

   In 2023, he coauthors “Unique Identification of 50,000+ Virtual Reality Users from Head and Hand Motion Data”. After training on five minutes of motion per person, the classification model identifies users among 55,541 participants with 94.33% accuracy from 100 seconds of motion.

8. 2025

   A Test of Time Award

   In 2025, he shares a SIGGRAPH Asia Test of Time Award with Tiantian Liu, Adam W. Bargteil and Ladislav Kavan for “Fast Simulation of Mass-Spring Systems”.

9. 2026

   Robo-Saber for game tests

   In 2026, he coauthors “Robo-Saber: Generating and Simulating Virtual Reality Players”. Trained on BOXRR-23 and applied to Beat Saber, the model generates VR headset and handheld controller movements from game object arrangements and examples of player styles.

10. As of October 2026

    Professor and company co-founder

    In October 2026, he is a professor in Berkeley's Division of Computer Science (EECS), with research in machine learning, image synthesis and image and video forensics. He is also a co-founder of KronosAI and Cosine Energy.