Research

After reading A Thousand Brains by Jeff Hawkins, I became deeply interested in cognition and the computational principles underlying intelligence. I joined the Ziv Williams Lab at Massachusetts General Hospital, where I studied language processing in the human brain. My work involves using natural language processing and machine learning to decode neural signals from intracranial recordings, with the goal of understanding how the brain represents and produces speech.

In high school, I worked at Lattice Automation where I developed ICOR (Improving Codon Optimization with Recurrent Neural Networks), a novel algorithm for optimizing gene expression in E. coli. The model used BiLSTM architectures achieving state-of-the-art performance, and was published in BMC Bioinformatics. I won 1st Place at Regeneron International Science and Engineering Fair for this research. I've also worked on medical imaging and clinical AI, including deep learning pipelines for auto-segmentation of cervical skeletal muscle in CT scans for sarcopenia analysis in head and neck cancer patients.

As a hobby, I've taken my interest from reading LessWrong to writing about AI bias. I've written a few pieces, and co-authored one collaborating with the Harvard AI Safety Team. I've also contributed to data projects such as Humanity's Last Exam.