Jun 2020 – Aug 2020 · 2 mos · India

• Worked on building a Virtual Yoga Trainer app that can detect real time yoga asanas.
• The pose is detected based on the angles extracted from the images using OpenPose.
• A dataset is created by extracting the angles(key joint points) from the images.
• The dataset was trained on different classification algorithms with different hyperparameters.
• A hybrid model of CNN and LSTM has been used to predict the user's asanas that outperformed the other models.
• An accuracy of 97% was attained of all machine learning models.

May 2018 – Jul 2018 · 2 mos · Mumbai Area, India

• Helped the web development team in updating the website with new content and design.
• Worked on both the frontend and backend part of development.

Jun 2016 – Jul 2016 · 1 mo · Mumbai Area, India

• Scraped news articles from the web and clustered a corpus of
• news articles into their respective domains using K-Means.
• Ran a similarity check to recognize the articles reporting on
• the same topic.
• Experience with performance analysis, optimizations and
• benchmark evaluations.