Disclaimer

Educational purpose

rlcapstone.ai documents personal learning projects. The models, apps, and results described here were built to explore machine-learning techniques end to end. They have not been clinically validated, independently audited, or certified by any regulatory body, and they are not offered as products or services.

Not medical advice

Some projects (such as MoleCheck) touch on health-related data. These projects are demonstrations of image-classification techniques, nothing more. They are not medical devices, they cannot diagnose any condition, and their outputs must never be used to make health decisions.

If you have any concern about a mole, skin change, or any other health matter, consult a dermatologist or physician. A model's "benign" output is not reassurance, and its "flagged" output is not a diagnosis — in both directions, only a qualified clinician can tell you what a lesion actually is.

Accuracy and limitations

Reported metrics (accuracy, sensitivity, AUC, and similar) describe performance on specific public research datasets under specific conditions. Real-world performance — different cameras, lighting, skin tones, and lesion types — can be substantially worse. Machine-learning models fail in unpredictable ways, and the write-ups here document those limitations deliberately, because understanding failure modes is part of the educational point.

No warranty

All content, code, and models are provided "as is", without warranty of any kind. Use of anything published here is at your own risk. The author accepts no liability for decisions made, or actions taken, based on this site's content.

Data and privacy

Projects on this site favor on-device inference: where a demo or app processes images, the processing happens locally and images are not uploaded to any server. Datasets used for training are public research datasets (for example, the ISIC Archive), used under their respective licenses.