Natalia Garrido

Developed machine learning and generative AI solutions on AWS, including LLM/RAG and model evaluation workflows. Designed MLOps and CI/CD pipelines for deployment, monitoring, and continuous model updates in production environments.

Built supervised and unsupervised machine learning models for classification, regression, segmentation, and predictive analytics. Performed EDA, data cleaning, feature engineering, model training/validation, and evaluation using standard performance metrics to support data-driven decision-making.

Integrated and transformed data from relational databases, Excel, ERP systems, and HR platforms to create analysis-ready datasets. Cleaned and normalized data with Power Query and built KPI dashboards in Power BI to support operational decision-making.

Activities and societies: AWS-based ML/GenAI development, MLOps/CI-CD, and model deployment/monitoring.

Currently working through a data science program focused on Python and data processing tools including Pandas, NumPy, SQL, and PySpark, with work involving AWS-based ML/GenAI deployment.

Earned an AWS Machine Learning Engineering certification focused on developing generative AI-based chatbots and applying advanced language model prompt engineering best practices to optimize user experience and response quality.

Earned the AWS AI Practitioner credential focused on designing, implementing, and evaluating predictive models using machine learning and deep learning techniques aligned with business objectives.

Activities and societies: Clustering and anomaly detection; EDA, data cleaning, feature engineering, and model training/validation.

Completed a certificate focused on unsupervised machine learning, including clustering and anomaly detection, with supervised and unsupervised model work such as segmentation and evaluation.

Completed a data scientist program including an Amazon Bedrock LLM solution to extract structured information from unstructured legal documents, using PDF/scanned text preprocessing to improve token usage and LLM input efficiency.

Activities and societies: Regression and classification; model training/validation; interpretation of results to support predictive modeling.

Completed a certificate focused on supervised machine learning using Python libraries such as NumPy and scikit-learn, including regression/classification and model evaluation with performance metrics.

Activities and societies: Multi-class classification with TensorFlow; RAG-based recommendations; predictive modeling for industrial electric motor failure risk.

Completed a certificate involving multi-class neural network development with TensorFlow and integrating Amazon Bedrock Knowledge Bases with a RAG architecture to generate technical recommendations.

Completed a quality engineering specialist program at Universidad Tecnológica Nacional in Buenos Aires.

Completed industrial engineering studies and developed a customer late-payment risk prediction project using an LSTM neural network, evaluated with AUC-ROC, recall, precision, and F1-score.