Ethan Kohrt

Programming Languages: Proficient in Python. Some experience in Java, JavaScript, SQL, and C#.

Tools: Scikit-learn, Keras, SpaCy, NLTK, Hugging Face, NumPy, Pandas, OpenRefine, D3.js, Excel.

Other – data visualization, regex, web scraping, Git, strong writing and presenting skills, Eagle Scout.

BERT Knowledge Extraction – extracting knowledge triples from BERT via auto-generated prompts. 

• Skills: hugging face, BERT finetuning, paraphrasing, implementing research papers.

Company Description Classifier – labeling company descriptions with industry tags.

• Skills: BERT sentence embeddings, sklearn, multiclass classification, web scraping.

Recipe Time Prediction – predicting the completion time of a recipe given its text. 

• Skills: sklearn, regression, grid search, natural language processing.

Chicago Police Data – investigating dataset of internal complaints about Chicago Police officers.

• Skills: SQL, data exploration, data visualization, sklearn.

Environmental Science QA System – prototype question-answering system for environmental topics

• Skills: QA systems, haystack pipeline, data collection.

See my portfolio page for code and many other projects.

Research Assistant: Hotel Recommender System

Furman University Computer Science, Summer 2020

• Implemented a recommender system for Walt Disney World hotels based on survey data provided by a travel agency.

IT Service Center Assistant

Furman University IT Service Center, January 2019 – March 2020 

• Consulted with customers over the phone and in person to understand and resolve problems.

Independent Study: Escape Room Generation

Furman University Computer Science, September – December 2019

• Pioneered methods of algorithmically generating structure and content for escape rooms.
• Investigated techniques such as TF-IDF and word embeddings, using spaCy and NLTK.

Research Assistant: Realistic Character Generation

Furman University Computer Science, Summer 2019

• Designed models for human-like values and emotions to procedurally generate realistic character behavior, for use in games and narrative generation. Showcased work at conference presentations.