• Leveraged knowledge in machine learning and data engineering to architect, execute, and lead the end-to-end development of a Serving Embeddings Platform.
• Maintained and enhanced an internal AutoML (Automated Machine Learning) system, introducing new features such as image classification support, to automate the entire process of creating a machine learning model.
• Conducted a Proof of Concept for Generative AI (LLMs) Tools.

Skills: Key-value store (DynamoDB), Go (golang), Google Cloud Pub/Sub, Jenkins, Kibana, Datadog, Python, Poetry, TensorFlow, Scikit-learn, SHAP, auto-sklearn, AutoKeras, PyTorch, CLIP, OpenAI API, LangChain, Hugging Face.

• Predicted customer's invested assets (Fixed income or variable income) using deep learning models for tabular data, resulting in a revenue of 15 million reais within a month, validated through A/B tests.
• Identified the group of the best representatives to handle a customer complaint using Natural Language Processing (NLP), leading to the company's best reputation score in three years.
• Performed customer segmentation based on business rules and machine learning models using a batch job with Airflow, resulting in a 10% increase in customer conversion for a segment above the user's original goal.
• Developed a web app for customer sentiment analysis and transformed machine learning models into APIs.
• Contributed to the design and development of an internal framework for the data science pipeline, covering cleaning, wrangling, visualization, modeling, deploying, and interpretation of data.
• Conducted evaluations of machine learning models.
• Conducted a proof-of-concept for predicting employee turnover using survival analysis and time series.

Skills: Python, SQL, Numpy, Pandas, Scikit-Learn, TensorFlow, NLTK, SciPy, Pytest, SageMaker, AWS Comprehend, S3, EC2, SNS, AWS Data Wrangler, Flask, Streamlit, TensorBoard, Pyspark, TPOT AutoML, SHAP Values, Docker, Airflow, MLflow, DVC, CML, Hugging Face.

Data Scientist 

• Conducted a Proof of Concept (POC) to detect clone events using a Siamese Neural Network and Nearest Neighbors.
• Conducted video content moderation using computer vision.

Main tools used: Python, Numpy, Pandas, TensorFlow, Keras, PyTorch, OpenCV, FAISS, AWS Rekognition, S3.

Junior Data Scientist 

• Utilized supervised machine learning techniques for content moderation, incorporating the detection of inappropriate words using Natural Language Processing (NLP) and images using computer vision.
• Designed and developed an internal framework for fraud event detection based on business rules and machine learning.
• Conducted the development of a web app for analyzing fraudulent transactions.
• Conducted model monitoring.

Skills: Python, SQL, Numpy, Pandas, Scikit-Learn, TensorFlow, Keras, OpenCV, Streamlit, AWS Rekognition, S3, Redshift, Docker, TPOT AutoML, Datadog.

• Developed the website for Control and Automation Engineering using HTML5, CSS3, and JavaScript.
• Actively participated in groups focused on continuous improvement.