Irina Krotova

• Solved business problems through deep learning, machine learning. • Trained and fine-tune models for NER and Intent Classifier chat-bot modules for optimal performance, achieved 2x error reduction • Led NLP module development for LegalTech tasks, including NER, Text Classification, Entity Linking, and OCR Errors Detection • Combined academic and industrial research experience developing LLM and other NLU models • Contributed to 5+ research papers published in top conferences such as ACL, LREC, and NLDB

• Designed and evaluated machine learning (ML) processes to automate NLP tasks for LegalTech. • Developed an OCR error detection module to improve pipeline performance by filtering out up to 10% of annotated data. • Created a text augmentation module, including NER augmentation, to enhance NER model accuracy and anonymize personal data. • Conduct exploratory data analysis, feature engineering, and data preprocessing.

• Developed experimental and analytic plans for machine learning algorithms and data modeling processes. • Implemented semantic clustering and topic modeling techniques for search queries for a legal search engine.