join our team
We are actively looking for individuals with deep learning expertise who want to work on new kinds of unsupervised and semi-supervised representation learning problems involving messy, heterogeneous, multi-modal, high-dimensional data from disparate sources.
Financial industry experience is not necessary. A PhD or master's degree in machine learning, computer science, or a related field is desirable but not required. We care more about your potential than your credentials.
We are located in the historic district of Old Town in Alexandria, Virginia, but are open to adding remote members to our team. We care more about having a great team than a geographically concentrated one.
How to Apply
Send an email to firstname.lastname@example.org with "deep learning expertise" as the subject. In the body of your email, besides introducing yourself, you must also answer at least two of these three questions:
Question 1: In deep LSTM networks (such as, say, Sutskever, Vinyals, and Le's architecture for sequence-to-sequence mapping), what happens inside an LSTM cell during a forward pass if the cell's forget gate has zero bias and a null weight matrix (or null weight matrices, if the gate's inputs are not concatenated)?
Question 2: In deep convolutional neural networks for image recognition (such as, say, a VGG ConvNet or one of He, Zhang, Ren, and Sun's deep convolutional "residual learning" networks), why are the learned kernel filters of every convolutional layer the same for each receptive field in the layer?
Question 3: How would you improve the second algorithm explained in this top-rated Quora answer? (Note: the author of that top-rated answer, Leo Polovets, is a member of our Executive Committee.)
The first two questions are trivial "FizzBuzz Questions" that anyone with deep learning expertise should be able to answer with ease, especially over email without time pressure. The third question is there so we can get a sense of how you approach and solve software development challenges.
As you write your answers, please keep in mind we will be evaluating, not just your understanding of basic concepts and issues, but also your ability to write about them coherently in everyday language. Avoid unnecessary jargon. And be prepared to explain, defend, and elaborate on your answers on the phone.
If we decide to follow up, we will respond to your email.