• Main project - analysis and improvement of the chatbot and the corresponding NLP Machine Learning model.
  • Performed exploratory analysis of chatbot conversations data. During that process found and fixed an error in database - 20% of classifier predictions haven’t been logged in the database because of invalid character limit set on a column.
  • Developed a system for monitoring the NLP machine learning model used in production and created detailed documentation in Confluence.
  • Built a Power BI report and presented the results to the management.
  • Retrained the model on new data (improved accuracy by 6%).
  • Created a Sankey diagram with Python that shows the flow of chatbot conversations.
  • Developed an NLP machine learning model to automate the manual classification of inputs the chatbot failed to answer.
  • Worked with Data Vault 2.0 methodology and ELT using Matillion ETL software.
  • Worked with Scrum using Jira.
  • Managed reports in Power BI Service and optimized their performance.
  • Testimonial