Natalia Cabeza

Computer Vision Engineer with hands-on experience developing deep learning pipelines for image analysis and computer vision applications. Experienced in designing, training and evaluating ML and DL models for classification and segmentation tasks using TensorFlow, PyTorch and OpenCV. Skilled in experiment design, model benchmarking, class imbalance handling and performance evaluation for real-world imaging problems.

Strong background in end-to-end ML workflows including data preprocessing, model training, validation and statistical analysis. Published researcher with hands-on experience adapting state-of-the-art architectures to domain-specific datasets.

Computer Vision Research Engineer at Connectivity and Signal Processing Research Group (CPS) - UIS (2021-11 – Present)

• Developed and optimized CNN-based computer vision pipelines for ultrasound and mammographic image analysis using TensorFlow, PyTorch, OpenCV and GPU hardware.
• Built end-to-end ML workflows including data ingestion, preprocessing, model training, evaluation and experiment reporting.
• Designed a deep learning pipeline for preterm birth prediction from first-trimester ultrasound images using CNN-based architectures.
• Conducted systematic model benchmarking and evaluation, including statistical validation, class imbalance analysis and performance metric comparison.
• Collaborated in the development of onboarding material and technical guidelines for ML evaluation and model assessment best practices.

Teaching Assistant at Universidad Industrial de Santander (UIS) (2024-02 – 2025-12)

• Teaching assistant providing support in signal processing courses and student learning activities.
• Completed institutional training in mediations of teaching and learning methodologies.

B.Sc. in Electronic Engineering – Universidad Industrial de Santander (2018-08 – 2023-12)

M.Sc. in Electronic Engineering – Universidad Industrial de Santander (2024-08 – 2026-07)