About Me
Hello, I'm Ansh Mehta. I'm currently pursuing my master's in Computer Science at Northeastern University in Boston.
Previously, I worked as a Software Engineer in the product analytics team at Whatfix.
My goal in pursuing this degree is to deepen my understanding of Machine Learning and AI. I'm actively enhancing my knowledge through coursework and self-study, while also implementing research papers as mini projects to develop a more profound understanding of the subject.
Project Demos
Semantic Product Search
Enter a product query to find semantically similar items based on their descriptions. (e.g., "summer dress", "running shoes for men").
It uses MiniLM-L6-v2 (Sentence Transformers) to generate vector embeddings and perform an approximate nearest neighbour search on the vector index.
Note: This demo is running on a CPU. Might be a little slow.
LSTM Classifier for Text
Classify news headlines into 4 categories: World, Sports, Business, Sci/Tech. Type at least 4 words. (Self Study in Spring 2024)
Checkout the project in detail on Github
Tetris AI (Deep Q-Learning)
A Tetris agent trained using Deep Q-Learning (Reinforcement Learning) to learn optimal gameplay strategies. Trained using an OpenAI Gym environment.
Checkout a demo below of the trained agent playing:
Checkout the project code on Github.