My company, Ozora Research, is developing a next-generation suite of Natural Language Processing tools centered around sentence parsing. We spend our time thinking deeply about the structure of language and then building statistical models that capture the structure. The requirements are: a good background in mathematics and statistics, and a deep interest in language and linguistics. Programming skill is a big plus, but not an absolute requirement.
I believe we are poised to make a huge dent in the world of NLP and linguistics. The field of automated parsing has been bogged down for the last 20 years because of its dependence on labelled training data (e.g the Penn Treebank). In my research I've discovered a way to circumvent that limitation, by building and evaluating sentence parsers using only a large amount of raw text data. I've spent the last couple of years validating the approach, and now that I'm confident it works, I'm looking to build up a team.
I have a nontraditional, minimalist and cockroachy business plan that won't appeal to everyone, but I can promise you that if I get rich you will also. I can also guarantee that you will get a ton of great experience that you would be tough to get at a big company. I want to especially encourage the following types of people to reach out:
- Recent grads in language-related fields who want to break into the software engineering industry
- People who have had some success in mainstream software engineering careers, but are looking to try something new
- People who want are thinking about going to graduate school in CS/ML/AI/NLP and want some experience doing research in that area
I'm happy to talk more about the specifics in person. Email me at daniel dot burfoot at gmail.