Theorem (www.theoremlp.com) is a company at the intersection of technology and finance. We were founded by a Google software engineer and a Morgan Stanley quant trader. We are a YC graduate (Winter 2014).
We are a cross-disciplinary team applying machine learning, software engineering and rigorous scientific investigation to revamp the lending and securitization space. Building good financial forecasting models is extremely challenging from both a technology and research point of view. Building models that you can rapidly test and see working in a multi-trillion dollar market is extremely satisfying.
We are a full-stack startup, not just another vendor; we make tools so that we can use them. We are profitable and have deployed over $100mm dollars. A large number of our investors are non-profits and university endowments, whom we are proud to be assisting. We previously raised a multimillion dollar seed round from top tier investors including major VCs, Max Levchin, SV Angel and Two Sigma.
Our problems are very technical: we are working with bleeding edge ML algorithms and feature engineering techniques. We work with approaches that span everything from traditional models like logistic regression, to SVMs, tree models and boosting, amongst many others.
We are a small team, with less than 10 engineers, based in Soma in San Francisco. We offer above market equity and cash compensation. We deeply value intellectual curiosity, creative idea generation, and close collaboration.
Interested? Email us at jobs@theoremlp.com
RESEARCHER
What you bring to the table:
* First and foremost, you should be a good scientist. You need to know when you have something and when you don’t, you cannot be fooled by randomness.
* You must be comfortable and enjoy working with data. This is the lifeblood of the business, your hypotheses start there and your testing ends there.
* You should be creative. You will be coming up with original hypotheses, and a framework to validate them.
* A high level of mathematical and statistical literacy, and an intense interest in applying quantitative analysis to do first class research.
* You must be comfortable writing code. Nobody will be doing it for you. The better you are at it, the more efficiently you can explore the model space and find winning ideas.
* We’re not just applying algorithms from existing libraries, we want a capable researcher who can understand our business problems and contribute to our analysis on a deeper level.
* Experience with feature engineering and mining external data to help boost model performance is useful.
* Research background devising novel algorithms and applying them to real world data to achieve robust, repeatable results.
* Familiarity with supervised-learning methods and modern statistics.
* Knowledge of quantitative finance is cool – but in no way a requirement.
What you’ll do:
* You’ll investigate data, construct a hypothesis, and think deeply about the results. After several iterations, you will discover something new. Instead of writing a paper, you will actually apply your ideas and see them working.
* You’ll need to care more about analysis being correct then achieving a p-value below 0.05.
* We’re exploring ideas from across computer science and statistics, including supervised learning, natural language processing, imbalanced data and anomaly detection, deep learning, time series, feature extraction and selection, and many, many other areas.
* We are very interested in ideas from biostatistics, survival analysis and epidemiology, so applied experience in these areas would be cool.
SOFTWARE ENGINEER
What you bring to the table:
* Professional experience in architecting systems, and with numerical or scientific computing.
* Experience with databases and dev ops preferred. Machine Learning experience is a big plus, but not required.
* Coding skill in Python, C++ or similar. We currently use Python, but welcome developers of any background, as long as you can pick up Python. Experience with numpy/scipy/pandas is a big plus.
* Experience in writing fast, performant code (especially numerical code) is a big plus. We operate in a competitive environment, and need our code to be fast. (No, we are not a high frequency trading shop.)
* Experience doing research in any scientific field is a plus.
* We value correctness, maintainability, elegance, and testability of code. We want to do things the right way over just getting things “done.” We’re strict about our code style and quality so that you don’t have to spend your time tracking down other peoples’ bugs.
* Knowledge of quantitative finance is cool – but in no way a requirement.
What you’ll do:
* We’re exploring ideas from across computer science and statistics, including supervised learning, imbalanced data and anomaly detection, ensemble learning deep learning and boosting.
* You’ll develop and implement solutions for real world, large-scale machine learning/statistical problems.
* You’ll build and maintain our best in class execution systems, data pipeline, backtester and model validation systems