Martin Espinoza

Data Analyst Machine Learning Engineer @Algocyte

Reading, GB

o Created interactive dashboards and detailed reports with Power BI, integrating data from multiple sources, applying Power Query transformations, and building DAX measures to track KPIs, effectively communicating key findings to stakeholders at all levels.o Developed Excel reports for review meetings, using dynamic charts, lookup functions,and pivot tables, with a professional layout to support decision-making.o Designed and trained Convolutional Neural Networks (CNNs) for image classification using Pytorch,integrating techniques like Transformers and Data Augmentation to improve modelgeneralization and accuracy.o Performed feature extraction from images and applied classification models, includingSupport Vector Machines (SVM), Random Forest, and Multi-Layer Perceptron (MLP),for effective data classification.o Applied dimensionality reduction algorithms such as PCA, t-SNE, and UMAP to reducedata dimensionality while preserving as much original variance as possible, supportingthe development of predictive models and facilitating the interpretation, exploration,and visualization of high-dimensional datasets.o Used Random Forest to assess feature importance, identifying key variables thatmeaningfully contributed to model performance and enhancing predictive accuracy.o Leveraged TensorBoard for hyperparameter tuning in CNNs using TensorFlow, streamlining themonitoring and optimization of the model\'s training process.o Applied Domain Adaptation techniques like CORAL to align feature distributions acrossdifferent datasets, enhancing the model\'s generalization to previously unseen data.o Developed predictive models using linear, polynomial, lasso and logistic regressiontechniques to estimate variables based on dataset characteristics.o Used Pearson’s correlation coefficient to assess the relationship between predictedand true variables, evaluating prediction accuracy and validating model outputs