👋 Hi, I’m Jonathan Stephens
I work in computer vision, 3D reconstruction, simulation, and robotics, with a focus on closing the Sim2Real gap.
I’m currently Chief Evangelist at Lightwheel.ai, helping teams use simulation, synthetic data, and world foundation models to train embodied AI systems. I also run Pixel Reconstruct, where I explore how machines perceive and model the physical world through topics like NeRFs, 3D Gaussian Splatting, and world models.
My interests include world foundation models, simulation-first robotics, synthetic data pipelines, and the future of physical AI. I share research, experiments, and tools through open-source projects, technical writing, and videos, and I collaborate closely with industry and research teams, including NVIDIA.
🛠️ Focus Areas
Computer Vision · 3D Reconstruction · Simulation · Robotics · World Models · Synthetic Data
🔗 Links
YouTube: https://www.youtube.com/@pixelreconstruct
Website: https://www.pixelreconstruct.com