M.S. Computer Science
Concentration in Machine Learning and Artificial Intelligence.
Colombia → Atlanta, GA
Daniel Duarte
Biomedical Engineer turned Computer Scientist, focused on medical imaging and machine learning
Colombian biomedical engineer currently pursuing a Master's in Computer Science at Georgia Tech (expected Dec 2026), with a background in biomedical technology and a growing specialty in AI/ML — deep learning, computer vision, and medical image processing in particular. Before the MS, I worked as a biomedical technician handling calibration, maintenance, and certification of medical equipment across hospitals and medical institutions.
Education
B.S. Biomedical Engineering
Experience
Biomedical Technician
Dec 2022 – Aug 2024
On-site installation, repair, calibration, and certification of medical equipment across hospitals and medical institutions.
- Managed installation, repair, maintenance, and testing of medical equipment, ensuring regulatory requirements were met and devices stayed properly maintained on routine cleaning/calibration/testing schedules.
- Tested and calibrated components against medical equipment manuals and troubleshooting procedures, producing timely and accurate service reports.
- Diagnosed defective equipment and provided technical assistance across dialysis clinics, endoscopy and plasma centers, urgent cares, research-aid facilities, and universities/schools — each with its own regulatory requirements.
- Maintained up-to-date training on hospital equipment including electrosurgical units (ESUs), automated external defibrillators (AEDs), IV pumps, and patient monitors, keeping documentation aligned with FDA regulations.
- Participated in depot repair training: disassembling and reassembling malfunctioning devices to diagnose, repair, and calibrate them.
ESUs
AEDs
IV pumps
Patient monitors
FDA compliance
Projects
CNN-Based Image Classification
Deep learning models using Convolutional Neural Networks for image classification and pattern recognition.
Developed state-of-the-art CNN models for image classification and pattern recognition tasks, focused on optimizing performance and accuracy using data augmentation, transfer learning, and regularization strategies.
Related certification: Nvidia's "Fundamentals of Deep Learning."
PythonPyTorchCNNsResNet
Generative-AI Diffusion Models
Explored diffusion models for image synthesis, style transfer, and denoising.
Worked with generative AI technologies, specifically diffusion models, gaining hands-on experience training and fine-tuning diffusion-based architectures for creative and scientific use cases.
Related certification: Nvidia's "Generative AI with Diffusion Models."
PythonPyTorch
AI-Based Anomaly Detection
Applied anomaly detection techniques across structured and unstructured datasets.
Applied AI techniques for anomaly detection using autoencoders, clustering algorithms, and statistical methods. Projects included fraud detection, network intrusion identification, and outlier detection in time-series data.
Related certification: Nvidia's "Applications of AI for Anomaly Detection."
PythonPyTorchscikit-learn
Tennis Ball Tracking & Court Analytics
Ball detection and tracking on broadcast footage, plus court keypoint detection and bounce in/out calling.
Trained on the public TrackNet dataset. Fine-tuned a YOLO11n model for ball detection, and built a custom TrackNet-style heatmap CNN in PyTorch as a stronger alternative for tracking a small, fast-moving object across frames. A separate YOLO pose model detects court keypoints, used to compute a homography from the camera view to a top-down court template so the ball's trajectory can be checked against the court lines to call bounces in or out.
PythonPyTorchUltralytics YOLOOpenCV
View code on GitHub →
Custom Image Filters
Custom image filter implementations for image processing tasks.
Designed and implemented custom image filters as part of broader image processing work.
PythonOpenCV
Search Algorithm Implementations
Implementations of classic search algorithms, including BFS, A*, and bidirectional UCS.
Implemented and compared classic search algorithms — breadth-first search, A* search, and bidirectional uniform-cost search — as part of coursework and personal exploration of search-based problem solving.
Python
Research
Medical Image Processing — PROMISE Research Group
Jan 2020 – Apr 2022 · Julio Garavito School of Engineering, Bogotá, Colombia
Investigated and implemented state-of-the-art image processing algorithms for medical applications, focused on brain lesion detection in tomography data.
- Used MATLAB's Image Processing and Wavelet toolboxes to pre-process tomography data, improving the sensitivity, specificity, and precision of segmentation and edge-detection algorithms used to detect and characterize brain lesions.
- Developed a suite of visualization tools in MATLAB to improve data interpretability and user experience.
AGoRa Lower-Limb Exoskeleton Prototype
Jan 2020 – Apr 2022 · Julio Garavito School of Engineering, Bogotá, Colombia
Contributed to a novel lower-limb exoskeleton prototype aimed at rehabilitating patients with hemiparesis.
- Normalized sensor data, ran tests, and recreated circuits for the exoskeleton.
- Carried out a literature review of state-of-the-art data acquisition technologies to inform hardware selection for the prototype.
- Performed hardware integration and testing, including circuit board printing and wiring for the exoskeleton's motors and microprocessors.
Skills & Certifications
Languages & Frameworks
Python (skillful)
MATLAB (skillful)
C# (beginner)
R (beginner)
Assembler (beginner)
ML / AI
PyTorch
TensorFlow
scikit-learn
CNNs
YOLO
ResNet
Diffusion models
Segmentation
Image classification
Feature extraction
Other
OpenCV
Data management & visualization
Spanish (native)
Certifications
- Fundamentals of Deep Learning — Nvidia
- Generative AI with Diffusion Models — Nvidia
- Applications of AI for Anomaly Detection — Nvidia
Get in touch
Open to opportunities in medical imaging, computer vision, and applied ML.