Kaitlyn Bartash '27: Optimizing AI Models for Liberty Mutual

These are the questions computer science major Kaitlyn Bartash ’27 has been digging into this summer as part of Liberty Mutual’s tech internship program.

For soon-to-be college graduates like Bartash preparing to enter a workforce fueled by AI, understanding how to use it as a tool isn’t just a skill, it’s becoming an expectation.

In Bartash’s internship, understanding is just the tip of the iceberg. In her role, she’s invested in a niche approach to working with AI called context engineering, aimed to improve the effectiveness and efficiency of the company’s AI models.

Focused on context engineering – delivering optimized data and information to an AI model to polish the results and workflows it delivers – Bartash is getting an inside look at how AI comes to conclusions and delivers results.

“If you’re looking at a 40-page book about animals, and the animal you’re looking for is on page 40, are you going to read the whole thing? No – AI operates the same when it’s scanning through prompts,” Bartash explained.

Over the course of the two-month internship, she has been contributing to a paper that will exist within Liberty Mutual about how context engineering works, as well as developing a walkthrough on AI ‘chunking’ methods. When these techniques are put into practice, they can help reduce hallucinations and help improve an AI model’s reasoning in multi-step prompts, like in operational workflows.

“I have such a better understanding of AI from being in this position and seeing how it works on the back end, [and] playing a role in how it’s best used and optimized,” Bartash said. 

Also as part of the internship program at Liberty Mutual, Bartash has been participating in discussions about responsible AI use, the tech market as a whole, and about entering the workforce and adapting from college to professional settings.

She recently participated the Liberty Intern Talent summit in Boston, which was a chance to connect with Liberty Mutual interns across the country, hear from speakers in the company and Bartash even had a chance to ask CEO Tim Sweeney a question.

The internship came at an opportune time, especially for someone who dove into computer science for the first time as a first-year at Emmanuel. 

Finding not only a great jumping-off point for a career post-graduation, Bartash is building experience and insight in one of the newest and most uncharted aspects of AI implementation in the workplace.

“[I’m] part of the first group of students coming out of college who are really equipped with this knowledge,” she said. “I’m realizing just how much AI use goes into everything.”

Emmanuel’s all-encompassing approach to computer science – including a focus on theory, ethics and a liberal-arts lens – has felt like a differentiator, Bartash said, putting her in a great position to explore complex questions about how to optimize AI use in a workplace.

“As one of the only liberal arts college represented here [in the program] I hold a level of pride being from Emmanuel,” she said.

In such a pivotal time for the growth of AI, it’s not lost on Bartash how much of an impact that she and the next generation can make by thinking critically about how to approach AI usage.

“I definitely am aware of the environmental implications of AI usage…it’s not going to go away – but the work I’m doing now can make an impact on that – by making the tool more efficient and using the correct prompting strategies, it uses less resources and costs less money,” she said.

Looking ahead to her senior year and graduation, Bartash is leaving her options open to see where her skills and insights can lead her for a career. She holds plenty of interest in AI overall, but would be open to any great opportunity across tech, programming or design – and even has thoughts about how she could explore AI implementation in UI and UX.