Himangi Mittal
Doctoral Student at Carnegie Mellon University
Pittsburgh, Pennsylvania, United States
Summary
👤
Senior🎓
Top SchoolHimangi Mittal is a PhD student in Robotics at Carnegie Mellon University specializing in physics-grounded generative models that internalize real-world material behavior. She developed a latent embedding that captures deformation and stress responses across elastic solids, sands, plastics, and fluids using large-scale distributed simulation training—work published at CVPR 2025. Currently she couples physics simulators like MPM with video diffusion models to produce visually compelling videos that obey real physics, blending simulator expertise with deep generative modeling. With a decade of experience spanning multimodal learning, large video-language and audio-visual models, point clouds, and autonomous driving, she codes daily in Taichi, Warp, PyTorch, TensorFlow, and C/C++. Her background includes research roles and internships at CMU and Honda Research Institute, and a track record of publications at CVPR, NeurIPS, and BMVC, reflecting both strong systems-level engineering and foundational ML research.
10 years of coding experience
2 years of employment as a software developer
Bachelor of Technology (with Honors), Bachelor of Technology (with Honors) at Jaypee Institute of Information Technology, NoidaCBSE Class X, CBSE Class X at Vishwa Bharati Public School, NoidaCBSE Class XII, CBSE Class XII at Ahlcon Public School, DelhiDoctor of Philosophy - PhD, Doctor of Philosophy - PhD at Carnegie Mellon University
Stackoverflow
29reputation
27kreached
1answer
0questions
Github Skills (70)
computer-graphics10
robotics10
rendering10
self-supervised-learning10
scene10
pointnet10
information-theory10
lidar10
optical-flow10
cluster-computing9
opengl9
kitti-dataset9
3d-reconstruction9
computer-vision9
gpu9
Programming languages (4)
Github contributions (5)
HimangiM/Just-Go-with-the-Flow-Self-Supervised-Scene-Flow-Estimation
Feb 2020 - Oct 2022
Self-supervised method for scene-flow estimation of LiDAR point clouds. Method is trained and tested on the nuScenes and KITTI datasets in TensorFlow. (CVPR 2020)
Contributions:88 commits, 9 pushes, 1 comment in 2 years 8 months
kittiscene-flowtensorflowmethodoptical-flow
HimangiM/RepLAI
Jul 2022 - Oct 2022
Contributions:56 commits, 6 pushes, 1 comment in 3 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.