I love to code π¨π»βπ»
π€ Humanoid AI Walker
- Developed a Humanoid AI agent that learned to walk by learning to control its 17 joints without any given data.
- Implemented the proximal policy optimization algorithm for the agent using PyTorch.
π Skim Unstructured Litrature Abstracts
- Unstructured abstracts can be hard to read. When the abstract contains more than 50 sentences, it is difficult to extract just the required information.
- Developed a web app that highlights only the sentence predicted as a userβs section of interest in such abstracts using neural networks.
- Reduced 90% of the time required to extract only the required information from such abstracts.
- Developed a REST API with CRUD functionalities that handles user registration and authentication, and allows users to create, view, edit, and delete blog posts.
- Programmed the API in python using FastAPI and Postgres database.
πͺ¨π§»βοΈ Rock Paper Scissor AI
- Developed an online rock paper scissors game to play against an active learning AI.
- Implemented the Markov model in JavaScript to predict a player's current move only based on their previous move.
- The AI will start to win more rounds by learning the players pattern by active learning, as the player plays more rounds. Try it out!
- Developed a Generative Adversarial Network (GAN) to copy the style of an image to another image.
- Programmed the GAN in python using pytorch.