Ayoosh Kathuria
Senior AI Engineer at Joist AI
Delhi, India
Summary
🤩
Rockstar🎓
Top SchoolAyoosh Kathuria is a Senior AI Engineer with nine years of experience building deep reinforcement learning and deep learning solutions that optimize large-scale industrial systems for energy efficiency while preserving safety and stability. Currently based in Delhi, he has progressed from research roles and applied R&D (including work on GANs for synthetic data at MathWorks and RL for autonomous driving at IIIT Delhi) to senior and lead engineering roles at AI startups and research firms. He combines hands-on model engineering—evident from PyTorch implementations of YOLOv3 and related tutorials—with production-minded RL research that targets real-world controls and power systems. Ayoosh publishes technical tutorials and blog posts (Paperspace) and has a track record of debugging CUDA/IoU edge cases and optimizing models (half-precision weight loads), showing attention to both reproducibility and deployment. He brings a practical blend of research rigor and product focus, often surfacing subtle implementation fixes that improve robustness in real-world pipelines.
9 years of coding experience
6 years of employment as a software developer
Summer School Student, Computer Science, Summer School Student, Computer Science at Indian Institute of Science (IISc)Bachelor of Technology (B.Tech.), Computer Science, Bachelor of Technology (B.Tech.), Computer Science at Shri Mata Vaishno Devi University
English, Hindi
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (11)
cuda10
object-detection10
computer-vision10
machine-learning10
pytorch10
python10
yolov410
faster-rcnn9
deep-learning9
mask-rcnn9
darknet8
Programming languages (4)
Github contributions (5)
ayooshkathuria/pytorch-yolo-v3
Apr 2018 - Jun 2018
A PyTorch implementation of the YOLO v3 object detection algorithm
Role in this project:
ML Engineer
ML EngineerContributions:53 commits, 2 PRs, 67 pushes in 2 months
Contributions summary:Ayoosh primarily worked on implementing and improving the YOLO v3 object detection algorithm in PyTorch. Their contributions included creating a function to generate test inputs, implementing the `parse_cfg` function for parsing configuration files, and adding layers such as `RouteLayer` and `ReOrgLayer` to the model architecture. They also added functions for loading and saving model weights, and optimized the code by converting it into half precision.
pytorchobject-detectiondetection-algorithmpytorch-implementationyolov3
ayooshkathuria/YOLO_v3_tutorial_from_scratch
Apr 2018 - Apr 2018
Accompanying code for Paperspace tutorial series "How to Implement YOLO v3 Object Detector from Scratch"
Role in this project:ML Engineer
ML EngineerContributions:22 commits, 27 pushes, 1 branch in 19 days
Contributions summary:Ayoosh primarily contributed to debugging and improving the YOLO v3 object detection implementation within the PyTorch framework. Their work focused on addressing issues related to CUDA compatibility, fixing bugs in the IoU calculation, and optimizing code for cases with no detections. They also added functionality, such as a command-line flag for video input and improving aspect ratio handling during image resizing.
object-detectorobject-detectiondetectorpaperspacescratch
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.