Leaving Watson Group - Saying Goodbye to a Job You Love

Wed, Nov 16, 2016 with tags jobs , watson , personal

About a week and a half ago I did something that a six months ago would’ve seemed nearly unthinkable. I left IBM, and more specificially, IBM Watson Group. It wasn’t easy, but now that I’m a few weeks past it, I’m know that it was the right move.

Perhaps the biggest challenge that I had when I was interviewing for positions was when each interview asked the inevitable question “So, why are you leaving IBM?”. My response was always, “I’m not leaving yet. I work with great people and I’ve got one of the coolest jobs in the world.” I still maintain that my position, was one of the coolest in the world, but with the successful launch of the [Watson Conversation Service] (https://www.ibm.com/watson/developercloud/conversation.html), a project for which I led development, I feel like the team is in good hands. I can’t wait to see all the amazing things they’re going to do in the future. However, the process of interviewing at different companies also opened my eyes to some of the imbalances you get from working at IBM. Toward the end I was having more days where I said “Adventure is out there” rather than “This is the best place for adventure”.

Some workplace processes were nearly constant at the places I interviewed - Slack, GitHub Enterprise, and AWS. Other things varied wildly. One place offered a phenomenal package, but had identified a single vendor, SAS, as their machine learning vendor. In today’s world, that’s a bad strategy for success. Another company proudly proclaimed that they had a 5000 node Hadoop cluster in one of their data centers - but when I asked what they did on it, they couldn’t provide much details and furthermore, many of their jobs weren’t well suited for Hadoop. Only a few places were looking at using GPUs for machine learning tasks - which is typically a good sign, especially if done properly. It seems like almost every company is trying to develop an AI and Machine Learning strategy, but most are still in the very early stages.

So, where does that leave me now? Well, I’m not quite ready to announce that to the world yet. However, those of you that know me won’t be surprised to find out that I’m working on building out an ML and AI cloud with bits of real-time interactions and nearly 100% full buzzword compliance. If you’re an academic friend, this doesn’t mean that I’m disappearing any more than I already had. I’ll still review papers and hope to occassionally show up to conferences again. With a little luck, I may even carve out some time to write some papers. Also, if you’ve got great people in software engineering, machine learning, data science, or design who are looking for a job, let me know. We’re looking in many different locations.