Aug 8, 2017 · 1 min read
It was interesting to me (a product/design/research consultant) see how you pull together estimates and work timeframes for a ML project. We have been doing some product and research on how a ML system should help people in a system, but haven’t built the actual models in house (we have partner consultancies that do).
How do you help identify the problems (or outcomes) they should apply to ML? How did the fictional client decide they needed “a document sorting system that gives BI analysts in a financial institution the ability to dump documents (PDF, Word, etc.) and sort them by semantic and syntactic similarity in many different ways?”
