Kudra

I already posted a question regarding a use case I have, where Kudra might give great value added. As I was very interested to see whether it might work out, I decided to jump the gun already and went Tier 2 to dip my toes into the water. Especially as I saw that they cited one other key player from my industry, Robert Half.

To give a short overview: I work as an IT recruiter and always anonymize the CVs of the candidates I work with to eliminate potential bias. This is usually a pain, as it involves meticulous searching to ensure that all personally identifying information has been removed and that nothing has been left in. As you can imagine, this causes me about as much joy as tax filing day and wisdom tooth extraction. But more often.

When I tried the pure Resume Parser function with my own CV as a test case though, I felt disappointed with the result. My CV is in German, as I operate from Germany. The classification accuracy was at about 40% - 50%. Names were mislabelled as skills, skills were mislabelled as locations or not recognized at all, company names were not found, locations were mislabelled... basically, at this stage, this would cause me even more work on the CV parsing side of things. As I use Zoho Recruit, I can compare it to the included CV parser of Zoho and have to admit that even though that one has nowhere near the features that kudra touts, it at least works correctly in 80% of the cases in labelling and extracting the relevant information.

I would really like kudra to work and be able to not only parse (again, that is the easy part) but concurrently be able to strip out or xxx out any identifying data. Actually, "like" does not correctly adress the emotion I would feel. Love kudra with the intensity of a thousand suns comes closer.

But at the current stage, I am sad and underwhelmed. Perhaps I did something wrong? Would be delighted to change the review if it helps me achieve what I am looking for.