Kaggle Competition: Explored self-testing & resource-constrained performance of Large Language Models on intricate science questions, trained using ChatGPT4.0 dataset. Offered insights into LLM potential & limitations in self-assessment.
Implemented CNN-RNN hybrid model using PyTorch for genre
classification of diverse music tracks. Leveraged CNNs and RNNs to
achieve accurate classification with over 80% prediction accuracy on new tracks. Preprocessed dataset of various genres. Input: grayscale MFCC spectrogram as numpy 2D array.
Transformed online shopping through AI innovation. Enabled users to engage in natural language conversations with ChatGPT, generating personalized product recommendations from Amazon's database based on distilled keywords.
A user-friendly healthcare recommendation website in Toronto, utilizing Cohere for symptom analysis and Google API to match users with personalized doctor recommendations based on symptoms. Streamlining the healthcare journey and enhancing accessibility.
Simulation on the Foot
Dec 2022- Apr 2023
Collaborated on foot nerve simulation machine project, supporting client's research. Researched foot nerve functions, developed
design solution using Arduino for pressure simulation and data
collection. Acted as primary contact, maintained regular communication and meetings. Tested actuators, prioritized safety protocols.
Garbage is a playful project featuring a useless trash bin that closes, laughs, and runs away when you attempt to throw something in. Built using an Arduino microcontroller and sensors, it adds a humorous element to the task of taking out the trash.
Journa Mobile App
Redesign of the Sidney Smith Building
Sep 2022- Dec 2022
Collaborated with a team to develop a comprehensive plan for the redesign of the Sidney Smith Building, with a focus on improving accessibility and social usability. Conducted research on user needs and preferences, and proposed design solutions based on findings.