I am passionate about ethical integration of machine learning into workplace environments and within governmental organizations. As a senior Computer Science student at Northeastern, I have recently co-oped with Northeastern's Center on Crime, Race, and Justice where I developed a classification algorithm to detect prison violence in 100000+ reports, giving crucial insights to Massachusetts prison staff about officer mental health and dramatically reducing workload for other researchers in the lab.

As a part-time research assistant mentored by Jessica Staddon, I have recently completed a project that utilizes Ensemble Large Language Model finetuning techniques to differentiate between scam and fraud narratives from Consumer Financial Protection Bureau data. The goal was to analyze which LLMs are most effective for detection and how narrative length and number of redactions effect performance. This resulted in the following paper that is under review: https://arxiv.org/abs/2412.08680.

As a Breakthrough Tech Artificial Intelligence Fellow, I applied the knowledge from my coursework as a machine learning consultant for the Bowerly Residential Community (BRC). We developed a chatbot that can aid overwhelmed night shift staffers easily access private corporate documents that outline procedures for emergency situations. As the BRC works to support New York's homeless population with housing and employment opporunities, there are a lot of crucial, time-sensitive situations where support is required. The following link redirects to the open-sourced project: https://github.com/ichadalavada/BRC_Nightstaff_Chatbot.