Summary
Akankshya Kar is a Senior Machine Learning Engineer with a decade of experience building production computer vision systems, blending research rigor from a CMU master’s with pragmatic product delivery at Motional and Apple. She has led mission-critical perception projects—introducing the first 3D image detection/segmentation for an autonomous platform, a traffic-light classifier with 90% F1, and a second-stage network that raised nighttime F1 by 10% while cutting false positives dramatically. Her work spans on-car multi-sensor tracking, active learning with generative vision-language models, and real-time on-device segmentation, and she has a track record of improving training throughput and reducing operational costs for cross-functional teams. Prior roles at Meta Reality Labs and Samsung produced advances in photo-realistic face avatars and an award-winning grain-quality app, reflecting a rare mix of deep learning research, embedded optimization, and product impact. Based in Santa Monica, she pairs strong C++/Python and PyTorch/TensorFlow skills with patents and publications, and she brings an inventor’s curiosity—she once performed as a professional Odissi dancer in the Commonwealth Games opening ceremony.
10 years of coding experience
5 years of employment as a software developer
High School Science with Maths and CS, High School Science with Maths and CS at Delhi Public School,NOIDA
Master's degree Computer Vision, Master's degree Computer Vision at Carnegie Mellon University
Bachelor of Technology - BTech Computer Engineering, Bachelor of Technology - BTech Computer Engineering at National Institute of Technology Karnataka
English, Hindi, Odia, Sanskrit