Summary
Pratham Mehta is a Machine Learning Engineer based in Bengaluru with a strong 3+ year track record designing and deploying computer vision and generative AI solutions across industry projects. He has hands-on experience improving object detection and segmentation pipelines—boosting YOLO V7 MAP50 from 89.4% to 97.8% and achieving mask mAP50 of 0.869—by integrating modern modules like CBAM, ConvNeXt and transformer-based techniques. His work spans 3D vision and pointcloud processing, diffusion models, LLM/RAG/VLM stacks, and robotics-focused AI, with practical edge deployment experience using NVIDIA DeepStream and AWS orchestration. Notably, he reduced annotation time by 30% via unsupervised methods and used Conditional GANs and self-supervised learning to significantly improve model robustness and accuracy. With an 11-year overall professional horizon and a BTech from CHAROTAR University, he blends research-driven experimentation with production-grade engineering. His GitHub (prathammehta16) reflects a practitioner who bridges cutting-edge papers and deployable systems.
11 years of coding experience
1 year of employment as a software developer
Bachelor of Technology - BTech, Bachelor of Technology - BTech at CHAROTAR UNIVERSITY OF SCIENCE AND TECHNOLOGY