Brett Martin. I wrote up this page as a way to summarize what these projects entail, and the next steps if you'd like to proceed.
First, some general information you might want to look through:
In general, the main features we look for in projects for this class are
- An interesting business problem that your company is facing. The problem should be well-defined-enough that students can start on it relatively quickly, but some amount of open-endedness doesn't hurt.
- A relevant dataset to go with the problem. This dataset should be relevant to the problem, relatively clean (though not necessarily perfect), and of medium scale (500MB - 2GB is perfect).
- A willingness to share this dataset with the students for the project. The students will be the only ones to have the dataset, and will sign basic NDAs (but nothing more complicated - in particular, we can't make the students sign any intellectual-property-related paperwork). The university, however, will not sign any agreement, and nor will the instructors. We will require all students to encrypt their hard-drives before they receive the data.
- An employee of your company who is willing to act as the main point of contact for the project. They will need to come to Columbia for the first class to present the project to students, and around 1 hour per week to answer students questions and/or meet with them.
We realize this is a lot to ask for. In return, we can promise a team of talented MBAs and engineers to work on your projects. Many teams have produced exceptional results over the years, some of which were later implemented at the companies they worked at, and many projects have let to job offers and hires from the hosting company. Note also that we will be supervising these teams closely - we usually have only 8 teams supervised by two instructors and two PhD TAs, so the students get a lot of help.
If this seems of interest, next steps would be a quick call to discuss the projects we've identified - reach out to let me know.