May 2020 – Jul 2020 · 2 mos · Bengaluru, Karnataka, India

• Developed various neural network models for image classification and time series prediction
• Created a time series prediction model for prediction of COVID-19 cases and deaths. This was done using NODE (Neural Ordinary Differential Equations). NODE is a family of deep neural networks proposed in 2018. Using NODE resulted in a drastic reduction of training time per epoch - existing LSTM models would take several seconds whereas NODE model took milliseconds
• Implemented transfer learning model using a convolutional neural network (VGG 16) and enhanced it by adding ODE (Ordinary Differential Equations) solvers. This resulted in an accuracy of 84.03% on corrupted CIFAR-10 dataset

Jun 2019 – Jul 2019 · 1 mo · Bengaluru, Karnataka, India

• Developed an Arduino-based Biobox to analyze gas concentration changes during photosynthesis
• Integrated oxygen, carbon dioxide, temperature, and humidity sensors for comprehensive data collection
• Utilized EEPROM to periodically save concentration values, enhancing data reliability
• Visualized results through serial monitor plotting, facilitating better understanding of photosynthetic processes