About

Prakash Chandra is an AI consultant with almost 9 years of experience in the field of computer vision, deep learning, and software development. He graduated from the prestigious Indian Institute of Science (IISc), Bengaluru, with a degree in Artificial Intelligence equivalent stream. He has been a major contributor to the Deep Learning with PyTorch Course by OpenCV, and is currently leading a deep learning team that is contributing to courses at OpenCV and blogs at LearnOpenCV. He has also created different pipelines that were needed to work efficiently as a team. In his previous role as a Research Engineer at Conduent, Prakash created a vehicle detection, deep-learning-based model using YOLO-v3 model, which led to 0.95 mAP at 0.5 IoU. He also improved the "Dynamic Hotspot (crime spot) Prediction Method and System" and filed a patent for the improvements. Additionally, he designed a loss function for number plate recognition, which ensures that any alphabet or number detected outside the number plate bounding box will penalize heavily, allowing for number plate detection and character recognition in one go. As an ML Software Developer at Amagi, Prakash and his team created a deep-learning-based scene change detection system and deployed it using AWS batch and Lambda function, reducing manual effort by 80%. He also did PoC for a deep-learning-based ad detection model and developed an ML ops pipeline for internal usage. He rewrote the algorithm part of an ad-scheduler, solving a numerical optimization problem with constraint. Prakash's education history includes a Master of Engineering (M.Eng.) in System Science and Automations from IISc and a Bachelor of Technology (B.Tech.) in Electrical Engineering from NIT Surat. His tech stack includes AI, Deep Learning, Computer Vision, OpenCV, Research Scientist, Software Engineer, Pytorch, ML, AWS, Terraform, Python, Flask, and Backend.