Joel Trulin

Designed a custom vision model for frame selection from camera, achieving 95% accuracy in detecting teeth aligners on a challenging dataset. Worked on transformer based multimodel like audio, text, vision capabilities.  PICOS elements & Entity Extraction using spacy & GPT models in health care.  Intelligent in selecting the optimal server for model training and inference with regard to cost efficiency.

Comprehensive experience in the entire Data Science lifecycle, from raw data cleaning and advanced data visualization and the development of cutting-edge machine learning models. Analyzing data and building statistical models using Neural Networks to solve specific business problems. Performing Classification, Regression, Time-Series Forecasting, Anomaly detection on Tabular, Textual, Image, Audio and Video data.