Amber Gupta
Data Scientist - Cv @Synchronoss Technologies
Signup · Get unlimited contacts
WORK HISTORY
Data Scientist - Cv @Synchronoss Technologies
Led development of I.P. Face Recognition pipeline using RetinaFace & ArcFace to replace AWS Rekognition APIResearched open-world object detection models for a text-to-image search, optimizing performance for CLIP & RAM modelsDesigned systems to auto-curate & personalize memories, flashbacks, and trip highlights to enhance user experience and engagement through intelligent content organizationCurated an object detection benchmarking dataset from OpenImages, Visual Genome, LVIS, V3Det, Object365 resulting in labels, a taxonomy & 10^6 imagesAchieved high/at-par metrics with in-house object detection & Face Tagging Pipelines against AWS Rekognition, bolstering client confidence from Verizon & AT & TDeveloped a context aware semantic text-to-image search, enabling dynamic recognition of concepts beyond simple objects
EDUCATION
University of Illinois Urbana-Champaign
Master's degree, Computer Science
The University of Alabama
Bachelor of Science - BS, Computer Science
SKILLS
ABOUT AMBER GUPTA
With nearly three years of experience as a Data Scientist at Synchronoss Technologies, I specialize in computer vision and data storytelling. I have led the development of innovative pipelines, including implementing a face recognition system using RetinaFace and ArcFace, and optimizing performance for text-to-image search models like CLIP and RAM. My focus has been on enhancing user engagement through intelligent content organization, such as personalized memories and flashbacks. My expertise extends to exploratory data analysis (EDA) and curating benchmarking datasets, having developed an extensive object detection dataset with over a million images and unique labels. I strive to deliver cutting-edge AI solutions, ensuring high performance and client confidence for enterprise partners like Verizon and AT & T.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.