|
Neuroscience-focused Biomedical Engineer Research Engineer in Computer Vision I work at the intersection of medical imaging, neuroscience, and computer vision. Most of what I build is about turning complex data into reliable workflows: segmentation and registration on DICOM images, EEG preprocessing and ERP / time-frequency analysis, and reproducible experiments that can actually be inspected, rerun, and extended. I care about interpretable systems, clean implementation, and research code that does not stop at a notebook screenshot. |
|
|
Focus AI for medical imaging |
Signals EEG / ERP / CWT pipelines |
Vision Segmentation / registration |
Workflow Reproducible, report-oriented |
|
What I do AI for medical imaging, neuro signal analysis, and computer vision pipelines with a strong preference for interpretable outputs over black-box polish. What I like building Reproducible workflows, clear evaluation setups, report-driven experiments, and technical systems that stay readable as they grow. Current directions Medical imaging analysis, EEG processing, representation learning, and computer vision problems where robust matching and structured evaluation matter. |
|
|
Domain MATLAB workflows for segmentation and rigid registration on DICOM-derived medical imaging data. Methods Chan-Vese and Malladi-Sethian level-set segmentation, plus NCC / SSD / Mutual Information registration metrics. Evidence A classical computer vision pipeline for biomedical analysis where segmentation and registration behavior stays inspectable end to end. |
|
|
Domain End-to-end EEG analysis in MATLAB, from preprocessing to ERP and time-frequency outputs. Methods Filtering, ICA, bad-channel handling, waveform analysis, grand averages, and CWT-based time-frequency exploration. Evidence Turns raw electrophysiology into analysis-ready signals and publication-ready figures inside one coherent, readable pipeline. |
|
|
Domain Neural network experiments across Python and MATLAB with reproducibility built into the workflow. Methods IF models, Hopfield networks, deep learning exercises, metrics, logs, and report-driven runs with a Poetry-managed setup. Evidence A compact record of repeatable AI experiments spanning dynamical models, classical networks, and deep learning exercises. |
|
AI and Representation Learning Experiment-driven work on embeddings, model behavior, and reproducible evaluation across AI workflows and neural network exercises. |
EEG and Neuro Signals Signal preprocessing, artifact handling, ERP analysis, and time-frequency methods for interpretable electrophysiology pipelines. |
Medical Imaging and Computer Vision DICOM-centered segmentation and registration workflows connecting biomedical engineering with classical and report-oriented computer vision methods. |
Open to research, biomedical engineering, and AI-for-health collaborations.
GitHub // https://github.com/DavideStefanelli97
Portfolio // {{LINK_PORTFOLIO}}
LinkedIn // {{LINK_LINKEDIN}}
Email // {{LINK_EMAIL}}
README calibration map
{{GITHUB_USERNAME}} = DavideStefanelli97
{{DISPLAY_NAME}} = Davide Stefanelli
{{PRIMARY_ROLE}} = Neuroscience-focused Biomedical Engineer
{{SECONDARY_ROLE}} = Research Engineer in Computer Vision
{{SUMMARY}} = Building AI-driven workflows for medical imaging, EEG analysis, and reproducible computer vision research.
{{LINK_PORTFOLIO}} = https://example.com
{{LINK_LINKEDIN}} = https://linkedin.com/in/your-handle
{{LINK_EMAIL}} = [email protected]
assets/profile-nameplate.svg, assets/head-mri-panel.gif, assets/freesurfer-panel.gif, and assets/signal-divider.gif are generated locally with python scripts/generate_assets.py.