From the archive.org cache (https://web.archive.org/web/20180824215739/https://www.opticsml.com/ https://web.archive.org/web/20180824215739/https://www.optic...

From the archive.org cache (https://web.archive.org/web/20180824215739/https://www.opticsml.com/ https://web.archive.org/web/20180824215739/https://www.optic...) of their homepage, it looks like they were a company that did OCR, indexing, & AI assisted data extraction. Seems like a pretty good use of ML really - shouldn’t be an intractable problem to identify something like a scanned W2, run OCR on it and extract the income fields.