cv
This is a description of the page. You can modify it in '_pages/cv.md'. You can also change or remove the top pdf download button.
Basics
| Name | Om Kumar |
| Label | Undergraduate Researcher in Machine Learning & Robotics |
| okhere21@gmail.com | |
| Phone | +91-8789603310 |
| Url | https://w-ok-e.github.io/ |
| Summary | Undergraduate student in Mathematics and Computing at MIT Manipal with experience in 3D vision, generative models, and autonomous systems. Worked on Gaussian Splatting, diffusion-based view synthesis, and medical imaging. Actively involved in robotics competitions and large-scale research projects. |
Work
-
2026.07 - Present Incoming Machine Learning Intern
Max Planck Institute for Intelligent Systems
Selected as a Machine Learning Intern for Summer 2026.
-
2025.05 - Present Summer Research Intern
NISER Bhubaneswar (Netweb Technologies)
Worked on 3D Gaussian Splatting and novel view synthesis.
- Developed post-hoc floater removal pipeline reducing Gaussian count to 20.4% while improving PSNR by +4.75 dB
- Built multi-view diffusion framework improving PSNR (+10%), SSIM (+15%), and LPIPS (-20%)
- Accelerated 3D reconstruction pipelines by ~1000× for TB-scale datasets
- Research submitted to ECCV 2026 (under review)
-
2024.11 - Present Research Intern
IIT Kharagpur (KLIV Lab)
Worked on medical imaging and super-resolution.
- Developed arbitrary-scale ultrasound super-resolution model outperforming baselines
- Built reproducible training pipeline with data versioning and experiment tracking
- Processed 100K+ medical images across distributed GPU clusters
- Working on multi-modal foundation model for medical imaging
-
2024.03 - 2024.09 Research Assistant
Manipal Institute of Technology
Applied Generative Flow Networks for drug discovery.
- Generated novel drug candidates for S. Aureus using GFlowNets
Education
-
2023.07 - Present Manipal, Karnataka, India
B.Tech
Manipal Institute of Technology
Mathematics and Computing
- Machine Learning
- Optimization
- Probability Theory
- Statistics
- Linear Algebra
- Real Analysis
Projects
-
Project Tirtha (Open Source)
Integrated VGGT-based reconstruction achieving ~1000× faster inference for 3D pipelines.
- Gaussian Splatting
- 3D Reconstruction
- Nerfstudio
-
Gurukul Platform
Full-stack tutoring platform enabling global onboarding.
- Next.js
- Supabase
- Scalable Web App
-
Monocular Depth Estimation for Terrain Mapping
Generated depth maps from UAV imagery for autonomous exploration.
- PyTorch
- OpenCV
- Robotics Integration
-
Object Detection & Localization (ODLC)
Real-time perception pipeline for autonomous vehicles.
- ROS2
- ONNX
- Embedded Inference
Skills
| Programming | |
| Python | |
| C++ | |
| Java |
| Frameworks & Tools | |
| PyTorch | |
| ROS2 | |
| Docker | |
| Git | |
| Gazebo | |
| MuJoCo |
| Domains | |
| Machine Learning | |
| Computer Vision | |
| Robotics | |
| 3D Vision |
Languages
| English | |
| Professional |
| Spanish | |
| Beginner |
Interests
| Sports & Activities | |
| Basketball | |
| Swimming |
| Creative | |
| Writing |
Awards
- 2025.01.01
3rd Place - IGVC 2025 AutoNav Challenge
IGVC
Achieved global 3rd place in autonomous ground vehicle competition.
- 2025.01.01
1st Place (Drones) & 3rd Overall - ISDC 2025
ISDC
Built monocular depth-based terrain mapping system.