As a machine learning engineer at Predata, you will work on our core algorithms for extracting volatility signals from raw sources, and their integration into analyst workflows for filtering through and understanding data.
You should be a insatiably intellectually curious. We’ll expect you to take projects from the idea stage to prototype to a finished product — this means proposing, prototyping, proving the utility of, and scaling and instrumenting both models and model evaluation techniques.
Our system takes thousands of discrete sources in every language, turns them into (sometimes nonlinear) time series (often exhibiting severe heteroskedasticity), and forecasts the likelihood of specific types of geopolitical events in times and places in the future.
We've barely scraped the surface of what can be done with our data, and are adding more sources and more data constantly. We've used or tried sparse regression, wavelet theory, proportional hazards models, and a laundry list of other things, but we're excited by the prospect of LSTMs or other Recurrent Neural Network architectures, language independent NLP techniques, ensemble methods for time series, graph-based learning, and other great ideas that pop up as we continue exploring our problem domain.
To apply, please send an email to email@example.com with: 1) Subject line ‘LAST NAME Machine Learning Engineer,’ 2) a résumé attached as a PDF. Please do not attach any other documents, such as reference letters or writing samples.