Abhijeet Mulgund landed the perfect internship after receiving almost two dozen rejections. As a computer science freshman at Rice, he didn’t expect to have immediate success when he began applying for internships after only two weeks of classes, but by April, he was feeling discouraged. He said, “Looking back, I can see that I really did not know what I wanted out of an internship, I just wanted to be an intern.”
In April, a family member working at Lexis Nexis heard about a new initiative: linking the use of natural language processing to search engines for legal data. The project manager was less interested in finding a programmer than a researcher, someone who could think about and discover commonalities in previously unrelated data models.
When Mulgund’s resume found its way to the project team, “they got really excited,” he said. “They called me, interviewed me, and hired me right away. Timing was everything.”
Rice’s CS program is known for teaching problem-solving before programming, which helped Mulgund discuss ways to explore the team’s proposed solution. He said, “They were just getting to the bleeding edge, working beyond their current areas of expertise. I’d done a lot of reading and taken a few online courses on machine learning. I knew what the algorithms were, could talk about areas they had not known much about.”
His Lexis Nexis mentor had an idea about using the same kind of algorithms that a company like Netflix might use to recommend movies to its users. But instead of recommending movie titles, he wanted to use the algorithms to find phrases or terms related to a specific document, or find related documents matching a set of terms.
“This summer, we are building models and figuring out how to make them work,” said Mulgund. “And we’re also learning what doesn’t translate from movie ratings to legal documents. The funny thing is, the structure you would use to train your model on the movie ratings data is the same for our data. It’s fascinating to take seemingly unrelated topics and find that, in reality, the underlying patterns are the same.”
Because his mentor gives him the opportunity to define what he’s going to work on, he feels he’s providing valuable input. Mulgund said, “I have the freedom to explore avenues that will most benefit the project and play off my existing skills.” But freedom to choose his direction doesn’t mean he can run around playing Pokemon Go. Mulgund is deeply invested in the research and hopes to publish a paper about it when they wrap up the project, just in time for him to return to Rice as a sophomore.
He also returns to Rice with a solid understanding of what he wants from his next internship. Mulgund said, “I don’t want my internship to be defined by the tasks I am given to do, but by the things I get to explore and learn. That’s why I’m having such a great time here. My mentor doesn’t say ‘do this, this, and this by Friday,’ he says ‘this is what we’re thinking’ and asks for my input.”