Flavio Loss

Currently my career goal is to work internationally in state-of-the-art Transformer models, as I follow specialization in the NLP field of study. I have almost 3 years of experience, and my skills range from building and deploying Transformers models to working with big data using Spark. I believe that AI can benefit a lot from engineers who understand text data and how Transformers works, and my current career path involves becoming a specialist in this field, creating responsible algorithms that generate business values for companies.

In my 2 years of experience as a Data Scientist, I have developed algorithms that process text data, leveraging the power of the Transformers architecture. In my first job at a startup based in Brazil,

we built a text data pipeline that given and input text, could find similar text in the company database. The algorithm for finding similar text uses pre-trained transformers models to find semantic similarity,

extract keywords and summarize the given input. The models used in this task were encoders like RoBERTa, Paraphrase and DistilBERT.

On the current company I'm working, we leveraged the use of decoders like Google's PaLM and GPT for finding useful information in patient medical records. The text data is provided as context to the model, and using a prompt we can extract critical information that could indicate a treatable disease.

During my undergraduate studies in engineering, I undertook an extensive research project focused on the classification of skin cancer. The research paper, titled "Deep and Machine Learning Approaches to Skin Cancer Classification Using NIR Spectrum Data of Skin Lesions," explored the application of advanced algorithms to analyze Near-Infrared (NIR) spectrum data collected from skin lesions.