Manushree Gangwar
Senior Machine Learning Engineer at Voxel51
Boulder, Colorado, United States
Summary
🤩
Rockstar🎓
Top SchoolManushree Gangwar is a Senior Machine Learning Engineer based in Boulder with a decade of experience building computer vision and deep learning systems for robotics, AR/VR, and autonomous vehicles. She has shipped production-ready pipelines for 3D object modeling, neural rendering and surface reflectance estimation, and has driven improvements from data capture hardware through photometric stereo to model training. Her recent work includes contributing evaluation tooling to the popular FiftyOne project to better surface COCO-style mAR metrics for object detection, and developing neural-rendering pipelines for close-to-real simulation at Torc Robotics. At companies from Magic Leap to Vicarious and Intrinsic she’s bridged research and engineering, turning NeRF research and BRDF estimation into robust synthetic training data for perception stacks. Known for improving data generation and debugging tooling, she often pairs model innovation with practical systems engineering to make perception work reliably in production.
10 years of coding experience
7 years of employment as a software developer
BITS Pilani, Birla Institute of Technology and ScienceMaster of Science (MS) Computer Science, Master of Science (MS) Computer Science at Columbia University
Github Skills (9)
object-detection10
computer-vision10
machine-learning10
python10
evaluation10
data-science10
metric10
deeplearning-ai9
deep-learning9
Programming languages (1)
Github contributions (5)
voxel51/fiftyone
Nov 2024 - Mar 2025
Refine high-quality datasets and visual AI models
Role in this project:
ML Engineer
ML EngineerContributions:44 reviews, 6 PRs, 16 pushes in 4 months
Contributions summary:Manushree made several contributions related to the evaluation of machine learning models within the FiftyOne framework. They added functionality to compute and display COCO-style mean average recall (mAR) metrics. The user also updated documentation and the model evaluation panel to reflect the new mAR features and enhanced the display of custom metrics. Their work focuses on improving the evaluation tools for object detection models.
pytorchpythonvisiondata-sciencedeep-learning
voxel51/inaturalist-2021
Feb 2025 - Mar 2025
Contributions:3 reviews, 5 PRs, 13 pushes in 29 days
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.