Computer field career?

How do I best transition into a career in computer vision research after a PhD in a different (but related) field?

• I am nearing the end of a PhD in computational neuroscience, with a focus on machine learning and statistical methods. My publication record is decent - good enough to get a good postdoc within my field, but no big hits. Some NIPS papers on a few different topics, basically. What would be the best way to transition into a career in more pure computer vision? I consider my coding chops good for a neuroscientist but I have no industry experience, and it's too late for an internship. Should I try to get a postdoc in a CS department? Try my luck applying for large tech companies with a research group? Go for a small startup with a lot of PhDs?

• Answer:

  The problem of entering the establishment, academia or large companies, at this stage would be that they probably won't let you diversify. A startup sounds like the best option but it doesn't have to have a lot of PhDs, someone has got to do the paperwork, in other words one group of too skilled people often has ego clashes with people not willing to be stuck with "mundane" things like taking care of cash flow.   Maybe you can find a backdoor into Computer Vision, I don't have much idea about computational neuroscience, but the word neuroscience makes me think about things like MRI, memory and how vision is perceived. Your expertise would be invaluable in making a system which can process static or dynamic images just like a human. With the military and the industry going big into drones, automation will be key and the drones will have to make sense of all the stuff their hi-resolution cameras keep streaming back to them. See if you can get an opportunity with people working in such areas, like Amazon.   And if all else fails, you can start off on your own, yes, funding can be an issue, so unless you find it or have sufficient motivation, you have to moonlight until you get your break.