Taro Logo

How to get ideas for patents?

Profile picture
Anonymous User at Taro Communitya year ago

None of my team mates have filed patents. Some Engineers working on ML, Computer Vision have filed lot of patents. I don't have experience in ML and its difficult to get into these teams as a beginner in ML. I am working in Insights team and building recommendation systems(Its in primitive stage, hardcoded rules). I have been trying to read lots of articles, what other companies are doing, recent ML work on recommendation systems but I am failing to come up with an idea that I can apply to our use case as I don't see similar use case outside. I don't have research engineers/scientists to talk to in my org. Filing a patent is not a requirement for my job role. I just want to do it out of interest. I am stuck. Could someone please provide insights on how I can move forward to think bigger?



  • 0
    Profile picture
    Robinhood, Meta, Course Hero, PayPal
    a year ago

    Given all this context, I highly recommend not going after patents. It doesn't help with career advancement, it takes up a lot of time, and most patents aren't the result of actually building something cool/useful. Check out the in-depth discussion here.

    Could someone please provide insights on how I can move forward to think bigger?

    When it comes to tech patents, generally what happens is big companies take some fairly standard tech they have built and use some language/lawyer tricks to claim it's unique enough to merit a patent. It's all quite the opposite of thinking big.

    If you have the time to chase after a patent, I highly recommend doing a side project instead.

  • 4

    I've found getting patents to be completely orthogonal to software work. The best way to learn how to do them at your company is to work on one. If you have engineers on your team who would be willing to have you help out with the patent, then you should definitely offer to do help as much as you can. In theory, this would be worth it for the experience alone, but it's possible they may decline (depending on how your company incentivizes patents and their personal incentives) since taking your help would almost certainly necessitate putting you on the patent. If it doesn't, then you could offer your help without being named an inventor since you're learning via this experience.

    While I agree with Alex that there's little reason for most software engineers to go after patents, if you choose to do so, that's the path I'd take: getting experience and mentorship by someone who has several patents. If you can't find someone willing to accept help with a patent, then learning about the ones your teammates put out and coming up with ideas could also be helpful. You can see what your teammates think of your patent ideas. They may help you refine them, give you a sense of direction, and possible suggest you go forward with an idea eventually (and they'll almost certainly help, so you'll get mentorship here too). Keep in mind that even if your patent is filed by Legal, it'll likely be 12-18 months before any conditional patent is available via public search and years before it's approved. And it'll be the company's patent, though you'll be listed as an inventor.

    Definitely consider your motivation before you go through this process though. It's a good amount of work that's unrelated to your work, and you won't advance your career this way (unless you're going somewhere where patents are the norm, e.g. some research scientist roles).