Harsh Maheshwari
Applied Scientist 2 at Amazon
United States
Summary
👤
Senior🎓
Top SchoolHarsh Maheshwari is an Applied Scientist at Amazon with eight years of hands-on experience applying deep learning and ML to production problems across computer vision, NLP and geospatial domains. A dual-degree graduate from IIT Kharagpur, he has built and deployed multiple low-latency, high-precision systems for advertising moderation, geocode prediction, and video auto-correction, including models optimized for Inferentia and large-model fine-tuning pipelines. He’s delivered measurable gains — e.g., a 23% improvement in address-completeness classification and recall/precision targets ≥95%/≥90% on moderation defects — while working end-to-end from data synthesis to deployment. Past research internships at Carnegie Mellon and published work in medical imaging and domain-adaptive detection highlight his ability to translate SOTA research into industrial products. Known for fast promotions and cross-functional leadership, he also brings practical expertise in LLM inference stacks, RAG, contrastive learning, and large-scale task-assignment simulation.
8 years of coding experience
4 years of employment as a software developer
Integrated (Masters + Bachelors) Electrical Engineering, Integrated (Masters + Bachelors) Electrical Engineering at Indian Institute of Technology, Kharagpur
Github Skills (45)
detach10
stacked9
unet-image-segmentation8
shift8
object-detection6
medical6
gradient6
temporal6
visual-recognition5
resnet5
densenet5
image-viewer5
pytorch5
notebook4
image-processing4
Programming languages (4)
Github contributions (5)
harsh-99/PyTorch-Tutorials
Mar 2019 - May 2019
Contributions:27 commits, 24 pushes, 1 branch in 2 months
harsh-99/Traffic-sign-detection
Aug 2018 - Apr 2019
Contributions:37 commits, 2 pushes in 8 months
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.