Systems and Computer Science Engineer at the National University of Colombia, deeply committed to academic excellence and professional growth. My passion lies in Artificial Intelligence and its subfields, including Machine Learning, Deep Learning, and Computer Vision. Additionally, I have a keen interest in Data Science and Big Data. I aim to develop innovative AI solutions that solve real-world problems.
-
AI Data Challenge (Source Meridian 2025): It was a team-based competition focused on building an end-to-end machine learning pipeline for real-world datasets. Our team achieved 2nd place overall, developing a robust classification system using XGBoost and automated data processing techniques.
-
Training Camp Medellín about Competitive Programming (EAFIT 2024): Intensive programming and algorithmic problem-solving camp designed to enhance computational thinking and efficiency in coding under pressure. Participated representing Universidad Nacional de Colombia, strengthening algorithmic and teamwork skills.
- B.S. in Systems Engineering and Informatics | Universidad Nacional de Colombia (In Course)
- Relevant Coursework: Data Structures, Algorithms, Machine Learning Fundamentals, Database Systems, and more...
Certifications
- [Deep Learning with PyTorch : Image Segmentation] - Coursera Project Network
- [Neural Networks and Deep Learning] - DeepLearning.AI
- [AWS Academy Graduate - AWS Academy Cloud Foundations] - Amazon Web Services Training and Certification
and more...
Check at all my certifications on Linkedin
Developed as part of the Special Academic Project at Universidad Nacional de Colombia, this initiative aimed to build an automated defect detection system for the metal sheets used in frying pan production at IMUSA. In collaboration with INTECOL S.A.S, and under the guidance of PhD. John W. Branch and Engineer Martín Aguilar Muñoz, we implemented a computer vision pipeline integrated with a SICK AI-powered industrial camera. The project evolved from a quality control challenge into a real-world AI deployment roadmap, covering data acquisition, model training, and system integration for industrial inspection. Beyond achieving a functional defect detection prototype, it provided valuable insights into scalability, real-time inference, and automation in manufacturing environments. Check it out here: LinkedIn post
[Detecting Tuberculosis in X-Rays - Datacamp Challenge] | [Python, Transfer Learning, Computer Vision, Optuna, MLFlow, Scikit-Learn, CNNs]
An applied deep learning project focused on detecting tuberculosis (TB) in chest X-ray images using CNNs and transfer learning (ResNet, DenseNet, EfficientNet). The work includes comprehensive data analysis, image quality assessment, and clinical-oriented metrics such as sensitivity, specificity, and predictive values. Designed as an AI-assisted screening tool, it demonstrates the potential of computer vision in medical diagnostics, emphasizing interpretability, model evaluation, and scalability for low-resource healthcare environments. Check it out here: Datacamp Notebook
[Medical AI Dashboard - XGBoost Literature Classification] | [Python, Next.js, Flask, V0, Scikit-Learn, XGBoost]
A data-centric project developed for an AI Data Challenge, focused on automating data preprocessing, feature engineering, and model evaluation for tabular datasets. The pipeline leverages Scikit-Learn and XGBoost to benchmark multiple models efficiently, visualize key metrics, and generate insights through a clean, interactive dashboard built with Next.js and Flask. Check it out here: Github repo
A modular multi-agent system that orchestrates specialized AI agents for marketing analytics, content strategy, and project management. The system features dynamic agent routing, collaborative workflows through shared memory, and a clean API interface. Each agent (AnalyticsWiz, ContentGenius, ProjectManager) specializes in different marketing domains and can access relevant data sources. The architecture supports traceability, reproducibility, and extensibility through prompt versioning and a well-organized codebase. Check it out here: Github repo
A data extraction tool that focuses on extracting explicit data from Commercial Real Estate Lease Agreements. In collaboration with @marco as a freelancer.
Planning on more stuff....
- Production and DevOps
- DL and ML
- Agents
"Colombia is home to the world's largest variety of butterflies, birds and second most biodiverse country on Earth, along with many other good things."
