Tech Lead

• Oct 2022 - Present

Senior Data Scientist

• Aug 2021 - Nov 2022

Senior Machine Learning Engineer

• Sep 2020 - Jul 2021

Data Scientist

• May 2019 - Aug 2020

● Responsible for building and deploying single-classification models to classify each article into a set of 30 different classes based on their topics or intent (e.g. Recommendation, Art, Hotels, etc). Built an hierarchical model structure with 3 different class tiers and a total of 6 classification models, all based on the text and title of an article. Achieved an improvement of 5%, reaching a 90% on the overall f-score. I was also able to improve the F-score for the most business relevant class by 25% reaching an F-Score of 85%. ● Built ranking model to get relevant articles related to a theme/keyword (e.g. list of most ”romantic” articles). Approach was based on similarity between embeddings where for each article a similarity score was assign. ●Developed a Text Generation solution with the goal to automatically generate/adapt description of items (e.g. Hotels, Attractions), in order to increase the number of available articles in the website and decrease the time need to create new articles (e.g. Top 10 Hotels in London). Responsible for exploring and testing Text Style Transfer and Data-to-Text approaches. [Python, Spacy, TensorFlow, PyTorch, LSTM, Transformers, SageMaker, SeldonCore, Kubernetes, FastText, StarSpace, Embeddings] Show less

Data Scientist

• Mar 2017 - Apr 2019

● Led and Developed a product that is able to extract, from unstructured text, structured information about acquisitions, partnerships, offers, headcounts and financial metrics, using NLP techniques. For example, information about the company that was acquired and who was the acquirer, how much that acquisition cost, etc. ● Achieved an accuracy of ~70% by analyzing the sentence structure and the grammatical dependencies of each word in the same sentence (semantic and syntactic analysis). [Java, OpenNLP , NLP] ● Co-organized the Artificial Intelligence Workshop on Nova University of Lisbon where we introduced the machine learning fundamentals. ● Part of the organization of Deep Learning with TensorFlow Workshop. The main goal was to introduce the Neural Networks and Deep Learning concepts using TensorFlow. [Python, Machine Learning, TensorFlow, DNN, Deep Learning] ● Mentored a junior element on the Sentiment Analysis library where an LSTM approach was implemented to classify text as positive, negative or neutral. ● Mentoring a junior element on PDF Table Detection component using a computer vision approach. Currently, researching a CNN network method. [Java, Python, Sentiment Analysis, Natural Language Processing, Machine Learning, TensorFlow, CNN, LSTM, Deep Learning] Show less

Data Scientist

• Mar 2014 - Feb 2017

● Developed an Opinion Mining system that analyzes opinions from unstructured information, using NLP and Machine Learning techniques, for Portuguese and English. The library is able to associate a sentiment to companies’ products and their characteristics. For example, the system extracts the overall market sentiment towards the new iPhone’s camera. ● The discovery of the products’ characteristics was done by studying grammatical relationships in a sentence. [Java, Python, OpenNLP, Scikit-Learn, Theano, NLP, Machine Learning, Feature Engineering, Word2Vec] ● Developer and maintainer of the search engine that analyses and transforms a natural language search query into a SQL query. In particular, I responsible for improving the engine lexical analysis phase. ● A new data structure and a lexical parser based on state machines was implemented which resulted in the improvement of the engine coverage (i.e. the system is now able to recognize more natural language queries). [Java, Clojure, NLP] Show less