Akamai interview experience @ Santa Clara

Interviewed at Akamai, Santa Clara a while back at two different teams. Position was Machine learning (ML) focused data scientist role (not a sql monkey role according to job description). This isn’t a 1st April joke. Team 1: Acquisition of Cyberfend Phone interview basic coding and no ML. I wanted to split on-site into 2 days for my convenience. Met 3 ppl. Coding test was easy. No one tested my ML skills because none of them knows any ML including the chief scientist who interviewed me. Hiring manager was a smart ass. Told me I have to work every other weekend and 3 days a week I have to stay back late. I decided not to come for the next round and told this reason to the recruiter. Team 2: Acquisition of Nomium Casual phone chat followed by on-site. No ML questions as no one knew any ML including their chief scientist. They just tested me on a badly written coding question that I solved easily and a Bayes’ theorem question that I solved without any issues. They told me they have developed something called domain2vec. When I asked how they did it, the person replied it is just few lines of simple Python code! Seriously what properties of domain are you trying to embed? Is it supervised or unsupervised like word2vec? And if you really understand learning representations, shouldn’t you be testing me on some basic ML? To top it off they told me they stopped using Spark as all their analytics could be done with Kafka. This made me convince that they are not into building ML products but just into analytics. I made it clear during interview that I will need spark to build ML products. They never got back to me afterwards.