cv

M.Sc

Universidad de Chile

Applied Mathematics

• Advanced Machine Learning
• Stochastic Simulatio
• Probability and Statistics in the Analysis of Data

Minor

Universidad de Chile

Computer Science

• Databases
• Algorithms and Data Structures
• Design and Programming Methodologies
• Software Engineering

B.Sc

Universidad de Chile

Mathematical Engineering

• Deep Learning
• Stochastic Calculus
• Nonlinear Optimization

AI/ML Developer

Bloom

As an AI/ML Developer, I specialize in the design, implementation, and optimization of end-to-end data solutions. My role focuses on building and maintaining robust and scalable MLOps pipelines on Google Cloud Platform, ensuring the performance and reliability of predictive models that improve efficiency in industrial processes for the water industry.

• Developed and implemented a predictive alert logic that minimized unplanned shutdowns and extended filter lifespan, while also leading the standardization of its technical documentation.
• Automated alert performance reporting (Kedro, Quarto), using time-series analysis to significantly reduce system false positives.
• Participated in the migration of ETLs to a serverless architecture on GCP (CI/CD, Docker), where I designed a code template that reduced development time by 60% and improved scalability.
• Automated ETL monitoring by developing a Cloud Function with Slack notifications, reducing incident response time from hours to minutes and eliminating manual supervision.
• Stacks: Data Science | GCP | Docker | CI/CD | Kedro | Quarto

Research Assistant

National Center for Artificial Intelligence (CENIA)

Developed a satellite image processing pipeline to analyze the evolution of wildfires in Chile, optimizing its efficiency and scalability.

• Analyzed wildfire data in Chile, extracting key metrics for segmentation.
• Automated the download and preprocessing of satellite images, improving data flow efficiency.
• Optimized queries in Google Earth Engine, accelerating them by 100x.
• Contributed improvements to the geetools library, benefiting the geospatial community.
• Stacks: Google Earth Engine | geopandas | geetools | scikit-learn

Research Assistant

Copenhagen Business School (CBS)

Researched and proposed new applications of the Bayesian Wasserstein Barycenter in the context of fairness models.

• Stacks: Fairness Models | Research | Wasserstein Barycenters

Master's Thesis Student

Center for Mathematical Modeling (CMM)

Calculated the Bayesian Wasserstein Barycenter (BWB) on a dataset of images using the Stochastic Gradient Descent algorithm on the Wasserstein space (WSGD).

• Implemented the WSGD algorithm abstractly, enabling its use by other researchers.
• Developed Wasserstein Generative Adversarial Networks (WGAN), optimizing training speed by a factor of 3x.
• Extended the WGAN architecture by incorporating an Autoencoder (AE), enabling a projector over the image manifold.
• Approximated the posterior distribution using GANs as the prior distribution, obtaining samples through the Markov Chain Monte Carlo (MCMC) algorithm.
• Developed a library to process and clean the Quick, Draw! dataset, implementing clustering algorithms to improve dataset quality.
• Stacks: PyTorch | WGAN | Autoencoders | MCMC | NUTS | Scikit-learn | Clustering | UMAP | TensorBoard

Freelance

Analysis of the Correlation between Inclusive Teaching Methodologies and Academic Performance of Primary Education Students

Conducted a statistical analysis to study the correlation between inclusive teaching methodologies and academic performance in primary education students.

• Used data analysis and visualization techniques to present results clearly.
• Applied statistical techniques to identify key patterns between teaching methodologies and academic performance.
• Performed categorical-numeric correlation analysis and hypothesis testing to validate the results.
• Developed a logistic regression model to predict academic performance.
• Stacks: Statistics | Hypothesis Testing | Logistic Regression | Pandas | Scikit-learn

Research Assistant

Center for Mathematical Modeling (CMM)

Performed feature engineering on seismic data and applied the Hawkes process for spatiotemporal probabilistic prediction of rock bursts.

• Developed processes to transform raw seismic data into relevant features, necessary for use in predictive models.
• Implemented the Hawkes process for spatiotemporal probabilistic predictions of rock bursts.
• Stacks: Stochastic Processes | Hawkes Processes | Feature Engineering

Data Scientist

Voyager Health

Predicted the INR index in patients undergoing oral anticoagulant treatment.

• Performed statistical analysis of INR time series, identifying key patterns.
• Implemented recurrent neural networks in TensorFlow and Keras, optimizing the training workflow.
• Stacks: Time Series | Recurrent Neural Networks | TensorFlow | Keras | Statistics

Freelance

Internet Service Allocation in Vulnerable Communities

Developed an optimization model to allocate Internet services in vulnerable communities.

• Formulated an optimal transport optimization model with capacity and demand constraints.
• Stacks: Optimization | Operations Research | Mathematical Programming | Python | PuLP

Data Scientist

SimpliRoute

Developed algorithms for anomaly detection in georeferenced data and service time analysis.

• Implemented clustering methods to identify anomalous georeferenced data, improving data quality.
• Developed an advanced system for detecting service times, enhancing logistics planning.
• Deployed key statistics for strategic analysis and decision-making.
• Stacks: Machine Learning | Clustering | Geospatial Analysis | Python | Scikit-learn | Pandas

R&D Engineer and Researcher Intern

SARCAN

Developed optimization models to improve route planning and efficiency in interurban transportation.

• Designed an action plan to enhance transportation operational efficiency. This plan is still being implemented today.
• Implemented optimization models to improve route planning.
• Stacks: Optimization | Operations Research | Mathematical Programming | Python | PuLP

Mentor in the AI Workshop for Students with Scratch

CONICYT

Taught students to understand, build, and use a Naive Bayes classifier in Scratch.

Mentor at Arduino Quest: Niñas PRO

Millennium Institute for Data Fundamentals (IMFD)

Guided students in developing an Arduino project.

• Participated in the 2020 and 2023 workshops.

Academic Team

Mathematics Championship (CMAT)

Prepared questions, graded exams, and answered inquiries during the competition.