Data Scientist | Experience in Model Development Lifecycle | AI Enthusiast | Building Innovative Data-Driven Solutions🚀

projects

hashnode

Shreyas Kulkarni  • Peerlist

Tools Used: Pandas, Modin, Scikit-Learn, Word2Vec.- Performed data preprocessing and EDA to extract meaningful insights.- Experimented with content-based and collaborative filtering methods.- Developed recommendation models utilizing TfidfVectorizer, Word2Vec, KNN, vectorization techniques, and cosine similarity.

People-to-People Recommendation Engine

<ul><li>Developed models based on Keras CNN and Huggingface ViT Transformer for classifying various table types and accurately identifying border types, and custom models for document categorization based on its unique types.</li><li>Built models to predict column lines, using Detectron2 utilizing annotation tools and image processing.</li><li>Contributed to developing NLP model using Spacy to extract specific elements from drug names</li><li>Developed automated pipeline for processing 50+ templates in healthcare/insurance domain, used primarily Camelot and Tabula for table data extraction, used Git for code delivery, and managing and addressing changes requested by clients.</li></ul>

Automated Formulary Parser for Insurance Documents

Collaborating with the authors to ensure technical accuracy and providing valuable insights to enhance the book's content.