1. Machine learning operations (MLOps)2. Face recognition: From noisy labeled datasets, building a backbone that can be used to solve both identification problem and verification problem.3. Face attributes classification: Classifying face images into multi-label and multi-class attributes.4. Object detection: Detecting tiny object in a high-resolution images and balancing trade-off between accuracy and performance.5. Generative Adversarial Networks (GANs): Solving… Show more 1. Machine learning operations (MLOps)2. Face recognition: From noisy labeled datasets, building a backbone that can be used to solve both identification problem and verification problem.3. Face attributes classification: Classifying face images into multi-label and multi-class attributes.4. Object detection: Detecting tiny object in a high-resolution images and balancing trade-off between accuracy and performance.5. Generative Adversarial Networks (GANs): Solving high-resolution images synthesis problem by using StyleGAN and controlling attributes of image generated by GANs.6. Video and text classification: Combining visual features from video and embedding features from text to classify. Show less

1. Generative adversarial networks (GANs): Solving high-resolution images-to-image translation problem.2. Noisy image classification: Classifying images which contain noisy pixels.

1. Python for Data Science2. Machine Learning3. Deep Learning

1. Organizing projects and events to support children following The Rights of Infants.2. Training and improving soft skills for volunteers.