Research Products
2025
Essay Evaluation Baseline
I trained several baseline BERT models for an AAAI iRAISE workshop paper in collaboration with Joey Hou and Dr. Lorraine Li. The models were trained to score essays given a particular prompt and criteria. I utilized starter code provided by the authors of the original paper. However, the starter code required extensive modification to make testable, including making a training pipeline, implementing their described custom loss function, and training the model for all essay prompts (as their example model was only trained for one of the dataset's prompts). Our paper and code are publically available.
2023
ELAN Data
A custom built API designed to create and work with ELAN files (.eaf extension) in Python. This library is open-source, and anyone can work on or contribute to it. Link.
APLS Automatic Speech Segmentation Pipeline
I utilized a finetuned pyannote model and custom built API to create ELAN transcriptions which were pre-segmented between speakers and phrases. Work is private and for the APLS lab. Email me for more information.
