Andres Aranguren

AI Software Engineer at Optiflux

Designed and implemented computer vision pipelines for fruit quality assessment, developing deep learning models for automatic segmentation, color calibration, and geometric measurement. Evaluated model performance under varying lighting and occlusion conditions, improving robustness for real-world agricultural environments. Deployed models to production and monitored performance using a reproducible CI/CD workflow.

AI Engineer at Deltaray BV

Developed computer vision models for automated defect detection using X-ray imaging within industrial inspection systems, contributing to both medical and automotive applications. Researched and implemented deep learning architectures for CT image reconstruction from incomplete projection data, improving inspection speed and reducing full-scan dependency. Investigated the integration of Vision Transformers (ViTs) into 3D convolutional architectures to enhance long-range spatial reasoning in defect detection. Collaborated on the development of a control mechanism linking the vision module with the X-ray acquisition hardware, enabling dynamic adjustment of imaging parameters based on model confidence and inspection results.

Data Scientist at Infometrika

Developed and deployed a computer vision system for the 2022 Colombian presidential elections to detect potential inconsistencies in official vote records. Designed an OCR-based pipeline to extract handwritten data from scanned electoral forms and validate results against public counting databases. Researched methods for improving OCR accuracy under real-world imaging conditions, including lighting variation, handwritten noise, and partial occlusion. Implemented model fine-tuning and post-processing strategies to enhance field-level extraction precision. Contributed to dataset curation and annotation strategies for large-scale document analysis, optimizing preprocessing workflows for thousands of heterogeneous voting record images.

MSc Computer Science Mentor at University of Leiden

Guided Master’s students in foundational programming, version control, and machine learning practices. Covered Git workflows, model evaluation with W&B, building ML portfolios, and reproducible research practices.

Fluid Mechanics Teaching Assistant at University of Leiden

Supported Transport Phenomena laboratory sessions, teaching fluid dynamics theory and finite element modeling using COMSOL Multiphysics. Supervised lab practices involving fluid flow characterization and Reynolds-number-based analysis.

Python and Computer Vision Teaching Assistant at Universidad de los Andes

Assisted in teaching Python programming applied to computer vision for a BSc-level Image Analysis course. Covered image processing basics, convolutions, kernels, and OpenCV introduction for real-world applications.

AI Software Engineer at Optiflux

Designed and implemented computer vision pipelines for fruit quality assessment, developing deep learning models for automatic segmentation, color calibration, and geometric measurement. Evaluated model performance under varying lighting and occlusion conditions, improving robustness for real-world agricultural environments. Implemented data preprocessing and augmentation strategies and deployed models to production with a reproducible CI/CD workflow.

AI Engineer at Deltaray BV

Developed computer vision models for automated defect detection using X-ray imaging within industrial inspection systems. Explored model interpretability and real-time control feedback for adaptive scanning. Researched CT image reconstruction from incomplete projection data and investigated integration of Vision Transformers into 3D architectures for enhanced defect detection.

Data Scientist at Infometrika

Developed and deployed a computer vision system to detect inconsistencies in official vote records during the 2022 Colombian presidential elections. Built an OCR-based pipeline to extract handwritten data and validate results against public databases; contributed to dataset curation and preprocessing workflows for large-scale document analysis.

MSc. Computer Science Mentor at University of Leiden

Guided Master’s students in foundational programming, version control, and machine learning practices. Mentored topics included Git workflows, model evaluation with W&B, building ML portfolios, and writing reproducible research papers.

Fluid Mechanics Teaching Assistant at University of Leiden

Supported Transport Phenomena laboratory sessions, taught fluid dynamics theory and finite element modeling using COMSOL Multiphysics. Supervised lab practices involving fluid flow characterization and experimental analysis based on Reynolds number and related flow regimes.