| Location: NYC
Remote: Yes
Willing to relocate: No I do bespoke clustering for data sets small or big (> Tb): images, text, sound, time-series, code. Most clustering methods break down when the data gets messy—overlapping groups, spherical clusters have too coarse resolution, unclear boundaries, weak signal. Tools like K-means assume simple shapes. Real data rarely behaves that way. I build custom clustering solutions that work on difficult datasets, where patterns are subtle but still there. I’ve applied this across biomedical data, legal documents, IoT sensors, imaging, and behavioral data. If your data has structure but standard methods aren’t finding it, I can help uncover it and go beyond k-means, cosine similarity and random forests. If an initial data analysis shows there’s no meaningful signal to extract, there’s no charge.
All clustering is done bespoke, reviewed by ML scientists. Email: "v" that works at @ https://denovoclustering.com/ |