At Zozimus we are always striving to stay ahead of the curve with ever-changing technology. With the recent emergence of AI, we are focused on following and implementing AI tools where and when we can for clients. This type of explorative work was demonstrated by summer 2023 intern Perry Sullivan, in the work he has been doing with Chat GPT and BARD. We got the opportunity to catch up with Perry to hear more about his time in the Digital Marketing department.
About Perry Sullivan
Perry, a rising senior at Boston University, says he “never really liked math but during my gap year I traveled, and I had an internship in Japan. That is when I started getting into cryptocurrency. You learn a lot from internships, but you can also learn a lot from teaching yourself, which is what I did. From 9am-7pm I was sending emails for my internship, but also gaining proficiency in spreadsheet work which is key to data science. Every night I went home and stayed up until 2am, teaching myself how to invest in crypto, which led me further into spreadsheet work. The crypto knowledge I gained turned into more research of the stock market. In my research of the stock market, I began to predict gains and losses. However, my time was soon interrupted by the COVID-19 pandemic. When I returned home from Japan, I continued to teach myself, and that is when I started dappling my personal knowledge into corporate finance. This is where my love for data science started.”
Discovering AI Tools
Perry officially started taking classes in data science a year and a half ago. His dad is a computer science teacher and finds that “data science is essentially computer science with a focus on data and numbers.” For Perry, the curriculum was brand new and that is when it really took off. Perry says he wants “people to use good data correctly. I’m not the show me the numbers guy because you don’t always know what those numbers mean. Knowing your audience and demographics is important and the data science aspect gives meaning to the numbers you put together.”
Perry was not expecting that he would get such a great opportunity to put these necessary skills to the test when it came to AI. However, when the opportunity presented itself, he was quick to jump at the opportunity.
“I didn’t expect to be working on AI during my internship. I was expecting to mainly be shadowing and doing spreadsheet work. However, my supervisors David (EVP, Digital Marketing) and Jake (EVP, Strategy) came to our group of interns asking for help with an insights project. After talking to them more, I was put in charge of researching AI. I started with prompt engineering, and then began trial-and-error work by creating a spreadsheet of prompts that would yield the best response.”
Perry assessed the responses he received from the prompt he provided the AI. These responses were ranked at levels 1-3 with three being the best desired response and one being the least desired. Each level is decided based on the amount of detail in the AI’s response. Perry noted that in this process he found “context is key to getting the most out of AI”, meaning the more specific information he provided the AI the closer he got to the desired response. He emphasized the specificity of the prompt you provided is of the utmost importance, because if there is too much information the response length grew shorter. The detail in the AI’s response increased when he referred to information, he provided the AI with earlier in the conversation.
ChatGPT and BARD
As for the differences between Chat GPT and BARD, Perry stated that the biggest difference between the two is that BARD can surf the web. This increase in available information makes it easier to provide context by being able to plug in links to articles the AI can read.
“Overall, I impressed myself through this research and look forward to delving deeper as I continue my studies of Data Science. Teaching AI to teach you is an example of what it is like to get a desired detailed response. You bring a car in for normal maintenance, so it doesn’t breakdown – same thing here with AI. I did not think I would be doing this but huge thanks to my supervisor David Wilson for getting me to dip my toes in this important work. I have now realized that this helps me think in a different way.”
Keep an eye out for Perry Sullivan. We hope to see him continue to make waves through the data science world as he continues his studies and moves toward his future career. Thank you for your work with us this summer Perry!
If you’re interested in being in our next class of interns, email your resume to email@example.com.