This was a biotech for the universal segmentation of human CT/MRI scans. I used a novel language model based routing strategy with a UNET model, where the encoder was enhanced by CLIP Embeddings. I worked on this with a group for a class called Lean Launchpad, in which another member and I focused on improving our model and building out our MVP. (Jan-March 2024)
For Treehacks 2023, my team and I made a TikTokesque ecommerce platform for reselling clothes. We used the OpenAI API to suggest a starting price based on the item's condition and a novel hybrid neural filter + knowledge graph convolutional network for recommendations. (Feb 2023)
I've worked on a few projects as part of Stanford's Association for Computing Machinery (Machine Learning Subgroup - MLab). For the first project I was a part of we looked at using small transformers to classify diseases and extract relevant population, intervention, and outcome data using scraped reddit threads. (Jan-March 2023)