European Union

Oct 2020 - Jan 2021

• Pre-processed EU’s text data in Python for optimising the search results of Watson Discovery Service
• Tested out-of-the-box techniques such as TF/iDF with various N-Grams and Words Embeddings for keywords generation that will improve the search accuracy of the delivered solution.
• Deployed the solution in IBM Cloud

Client under NDA

Mar 2020 - Jul 2020

ABN AMRO

Oct 2019 - Dec 2019

• Responsible for designing and devloping an enterprise-level unsupervised learning solution for the bank’s chatbot intent classification. Solution was deployed in AWS and it is been used bi-weekly to discover new client intents.
• Hyper-parameter tuning among clustering algorithms such as Gaussian Mixture Models, LDA and NMF to determine the optimal for discovering new user intents
• Exploratory data analysis of the user logs in order to understand better the data beforing feeding them to the algorithms

NATO

May 2019 - Oct 2019

• Data wrangling and analysis of the received text data (job descriptions)
• Worked with spaCy library for natural language processing
• Used stemming and tokenization pre-processing steps for data quality improvement