Software Is Eating The World

Machine Learner

San Francisco, CA, US

Apply for this job

We are UnifyID, the TechCrunch Disrupt 2016 runner-up company that is building a revolutionary identity platform based on implicit authentication. Our solution allows people to identify themselves in a unique way that is extremely difficult to forge or crack. Best of all, we are doing it in a way that respects user privacy.

We are looking for a math+code engineer/signal-processor/hacker/self-proclaimed guru who is comfortable with crafting, hacking, implementing, re-implementing and most importantly, breaking Machine Learning algorithms deep, shallow or otherwise.

If you think you have the answers/(mis-)informed opinions on one or more of the following questions, we'd like to hear from you! (a.k.a. you'll love it here!)

1: The great generative model wars [GANs vs Pixel-RNNs vs VAEs]: Who do you think will win and why?

2: Word on the Kaggle street is that XGBoost is killing it! Why do you think this is?

3: Do you think that the problem of counting independent sets in a bipartite graph is not #P-complete but #BIS? Why so?

As language/platform agnostic as we are (we use Lua, Python, Julia and R on a daily basis and are eagerly awaiting the Milk compiler from the CSAIL folks too), we expect you to be unreasonably good at and evangelize at least one of the tuples in the Cartesian product of L X P X O where L={Python, Scala, Julia, R, Lua, C++, Java}, P={Scikit-learn, Torch/Autograd, Caffe, Keras with Theano/TensorFlow, Chainer}, and O={Ubuntu, OS X, RHEL / CentOS / Fedora} and pick things up when required.

If you think this might be a good fit for you, drop us a message with your favorite moment in the sun (publication, GitHub code-base, live-project link) and answers to at least one of the three questions above.