- Although AI-related jobs are booming, many universities still do not offer specialized AI degrees.
- Allison Krinsky, a data scientist at JPMorgan, values hands-on experience over formal education.
- Krinsky encourages people to build their own AI projects, such as a travel system or a sentiment analysis model.
AI-related careers are popular right now, and many schools are adapting their curricula to include AI-related coursework, but there are still only a few schools that offer AI majors.
Allison Krinsky graduated from the University of Washington in 2022 with a degree in computer science. She currently works as a data scientist at JPMorgan and makes videos about tech careers in her spare time. She told Business Insider that many majors are interchangeable and that some degrees, such as computer science, mathematics, computer science, and data science, could lead to jobs in the field.
But even though Krinsky studied a traditional curriculum that prepared her for a technical career, she says it was her work in the lab that has advanced her career more than anything else: During her year there, she says, she did “a ton” of things, including building models and managing databases.
Most AI-related jobs require a technical portion of the interview process, and applicants need to be able to talk about projects they've worked on, Krinsky said.
“In my interviews, I'm often asked questions about what I've built, what I've done, what problems I've faced,” Krinsky said.
Krinsky said that while big tech names might look flashy on a resume, work experience is key to actually getting a job, and that his internship before joining the lab involved smaller projects that didn't require a lot of skill.
“An internship is a great opportunity to say someone hired me, and it gives you a little bit more credibility,” Krinsky says, “but if you haven't had a traditional internship, it doesn't put you out of the running.”
As demand for AI-related jobs grows, some companies are becoming more strict about the talent they're looking for, so if you have limited experience or want to beef up your resume, it's not a bad idea to start your own project to hone your skills. Depending on the type of job you're interested in, Krinski said there are a few paths you can take.
One option Krinski recommends is a travel recommendation system built using a large language model, a project he says could be carried out in a variety of ways, even with limited experience, including prompt engineering, search expansion generation and fine-tuning.
Krinsky also proposed using natural language processing to create a sentiment classification system for reviews. This involves extracting information from text data and categorizing it into entities such as positive and negative sentiment, he said. Krinsky said this could be used for financial analysis and identifying investment opportunities and risks.
Krinsky said people could also try their hand at image recognition or computer vision projects, which involve finding a set of labeled photos and teaching a computer what's in the images, which she said is a good way to learn about neural networks.
Krinski says these projects can take anywhere from one to three months, depending on how much free time you have. Most projects start with scraping data from the web, then require building, training, and fine-tuning a model. Krinski also recommends writing a report detailing the project process and results to demonstrate your work.
Projects don't have to be revolutionary, she said, but they should experiment with multiple datasets and be able to explain what's going on. Anyone can reproduce the code in the tutorials, so it's important to add your own unique aspects, she said.
“You have to get rid of the notion that, 'I just wrote the code and nothing went wrong,'” Krinsky said.
