About

Rohit Choudhary is an experienced Machine Learning and Deep Learning Engineer with a strong background in Transformer models and attention mechanisms. He is skilled in using TensorFlow, PyTorch, and a variety of programming languages, including Python and C++. Rohit has 3.69 years of relevant experience working in a startup environment, and he is comfortable with fast-paced, iterative development. Currently, Rohit works as a Machine Learning and Computer Vision Engineer at shoefitter, where he has developed a 3D reconstruction solution using Point Cloud Feature extraction and Registration, SLAM Pose Graph, Visual odometry, and Linear optimization. He has also written microservices on the NestJs framework, deployed on an Azure Kubernetes Cluster, and implemented Github action/workflow for CI/CD and Argo CD for deployment. Additionally, Rohit has developed a Javascript widget using Angular, which connects to deployed microservices through a GraphQL API and is served by Nginx. Before joining shoefitter, Rohit worked as a Research Assistant at the University of Konstanz. Here, he worked on Light Field Synthesis using a Recurrent Neural Network of 64 time steps. He trained the network on a synthetic dataset and was able to extrapolate given input views with low angular and spatial resolution to a higher resolution. Rohit used Attention, Transfer Learning, and Auto-Encoders to achieve the final results. He also worked on the detection, tracking, and visualization of insects under controlled light settings, using Python, OpenCV, and TensorFlow. Rohit holds a Master's degree in Informatics from the University of Konstanz and a Bachelor of Technology (B.Tech.) in Information Technology from IIIT. He is proficient in Computer Vision, ML, Deep Learning, NLP, Neural Networks, and OpenCV. He also has experience with CI/CD, Kubernetes, Azure, AngularJS, and microservices.