- Programming Languages: Python, C/C++, MATLAB, HTML, CSS, SQL
- Libraries: PyTorch, TensorFlow, Keras, scikit-learn, Light GBM, XGBoost, pandas, Matplotlib, seaborn, Plotly, Dash, OpenCV, BeautifulSoup
- Modelling: CNN, Transformers, Regression, Classification, Clustering, Time-Series Analysis
- Software & Tools: MS Excel, VS Code, LAMP Stack, Git
B.S., Electronics Engineering | University of the Philippines Diliman (July 2023) | Summa Cum Laude (GWA: 1.1378) | DOST-Merit Scholar
- Relevant Courses: Machine Learning, Probability and Statistics, Digital Signal Processing, and Data Structures and Algorithms
- Activities and Societies: UP Digital Signal Processing Laboratory
Data Scientist Intern @ UnionBank of the Philippines (April 2023 - September 2023)
- Contributed to the development of the Explainable and Responsible Artificial Intelligence (XRAI) guidelines and toolkit, providing data scientists with a better understanding of the models they build.
- Authored a new section on the XRAI guidelines about the feature selection techniques and some AI-related risks to look out for.
- Developed a comprehensive and easy-to-understand dashboard for the entire toolkit, providing an overview of the results obtained from evaluating the model on each available tool within the XRAI toolkit.
- Created a new feature for the toolkit by leveraging the AIF 360 fairness metrics and bias mitigation algorithms, including it to the wide-array of tools of XRAI.
- Deployed the toolkit for testing with other data science teams to evaluate its capabilities and gather feedback.
Multi-Stage Hybrid-CNN Transformer Model for Human Intent-Prediction (August 2022 - July 2023) (Repository)
- Led a team of 4 in designing and implementing a model that can predict human intention based on the objects they are gazing at using computer vision techniques.
- Developed a multi-stage model that includes a pre-trained monocular depth estimator, a Transformer intent classifier, and a CNN-transformer gazed object predictor using Python.
- Achieved an overall accuracy of 54%, precision of 46.19%, recall of 50.01%, and an F1 score of 45.53%, demonstrating the model's robust performance in predicting human intention.
- Accepted in 2023 IEEE Region 10 Conference (TENCON) to present the paper for publishing.
Polynomial Regression Model (November 2022 - December 2022) (Repository)
- Built a polynomial regression model using Python that can solve any polynomial equation by outputting the degree and coefficients, given a set of x-y pair inputs.
- Implemented the Tinygrad framework and SGD optimizer to find the optimal degree and coefficients for the model.
- Achieved a perfect R2 score of 100% and an RMSE loss of 9.56.
- Showcased skills in Python programming, mathematical modeling, and optimization techniques.
Projective Distortion Removal (October 2022 - November 2022) (Repository)
- Developed a python program that removes the inherent projective distortion on a rectangular object in an input image.
- Solved for the homography matrix given the four corner points of the rectangular object and performed linear transformation of the image using this matrix.
- Utilized OpenCV library for GUI design.
EMG-Based System (February 2022 - June 2022)
- Designed and implemented an open-loop EMG-based system consisting of an analog front-end unit, a transceiver unit, and a signal processing unit.
- Led a team of 3 in designing the analog front-end circuit and the transmitter side of the transceiver unit.
- Presented the design considerations and results to the project advisor.
- Utilized TINA for circuit design, MPLAB for the ADC of the transmitter using C programming language, and MATLAB for digital signal processing.