Job Description Draft

The Summer Intern

As a child, I often woke up to my dad kissing me goodbye before leaving for work. My half open eyes would see him tuck in his formal shirt and fasten one of many belts he had. I had always wondered what it was like to be in his place. To drive a car to office, have access to a computer, chat with fellow workers, and have an unlimited supply of snacks! All he had to do was sit around, punch in a few numbers in his computer, and he would be minting money! How simple right? No. Work is like Fruit Ninja. You are always slicing away task after task. It never ends. Pressure surrounds you from all directions. Whether it be from your boss or meeting deadlines after deadlines. I slowly started to understand that the work my father did, or as a matter of fact any job, is not always a low hanging fruit. This realization hit me when I started working myself. During my 6-weeks as a summer intern at an education technology startup in Seoul, South Korea, I had multiple responsibilities. If you have interned before, you would know that traditionally an intern is more of an assistant role where you help out with office tasks and learn from your supervisors. My experience went beyond this, allowing me to get my hands messy. To give you a better understanding of my major duties, I have highlighted three descriptions that I will explain with examples: 

• The Scientist
• The Builder
• The Author

The Scientist

As a scientist during the first phase of my internship, I had to put on my curiosity hat and do the following things: 

• Play around with the various features of the educational application and create a user manual
• Interview users of the application
• Experiment with Artificial intelligence models like ChatGPT

This initial phase was the least challenging but required an open mindset and willingness to learn. I first had to learn about the company's product. I opened my phone and spent hours trying every single feature and mode of their app. In fact, I was so invested that I had encountered several bugs (design/software defects) and reported them to the corresponding team heads. While navigating through the app and learning about each feature, my supervisor recommended me to create a user manual, to which I did. Next, I would also need to learn about the users of the app and see their perspectives. Through virtual chats, I interviewed some of the app users, primarily students and teachers, who shared with me their feedback. This process gave me a better understanding of what worked well for the users and what didn't. Then, I decided to pitch an idea to the CEO of the startup to include an AI feature in the app that would help many of the student users. To my surprise the CEO agreed and I immediately began researching and experimenting with AI models like ChatGPT and Perplexity to see which one was suitable to integrate in our app. Me and another summer intern played around with these models to test out which prompts would generate our desired response (This process is called prompt engineering). After learning about the product, collecting responses from users, and testing out which prompts worked best with each AI model, it was time to get into the real action.

The Builder

In this role, I had the chance to actual apply my experimentation results and turn my idea into a tangible prototype. I accomplished this by doing these three things: 

• Outlined and drew a user flow with specific details for each stage of the AI feature
• Created a visual representation on a software tool of the AI feature
• Worked with backend software team to implement the desired AI model in the app

When building something, you need to have a blueprint. When I built my prototype, I drew out a user flow on several pieces of paper to map out how the user would activate and navigate through the feature I was about to build for the app. I also wrote down specific details, functioning, and purpose for each stage of the feature. For instance, when a student is solving problems on the app and needs to quickly activate the chatbot, they can shake their phone and the chatbot would appear on their screen. Then, I took this rough blueprint and learned how to convert it into a visual representation on my computer. This is because the computer tool that I utilized could show exactly how the feature would look like visually before it would be released. So, I took my time thoroughly choosing the layout, color scheme, and even smaller details like the shape of the text bubbles for the chat screen. I then handed over this design to the backend team who brought the design to life in the app. Finally, I worked alongside senior software developers to integrate our chosen AI model. I did not have the technical expertise to actually code, but I sat next to the software developer who showed me the whole process over a span of 3 days. We had to keep editing the code as the AI model was acting weird. At last, I had successfully designed and implemented my idea. For the first few weeks, we only released the beta version, and monitored the response. 

The Author

For the last 2 weeks of my summer internship, my duties included: 

• Researching about technology-based solutions in classrooms
• Writing a paper that proposes a solution

The edtech startup I interned at had more than just an app. They also had physical clickers that they sold to classrooms. These clickers were powered by bluetooth and students can use them to answer quiz-style, multiple choice questions that the teacher would project on their screen. My task was to scour the web for any research on the effectiveness of classroom response systems like these clickers. I ended up finding many articles and scientific studies that supported the use of these technology based solutions in educational settings. After collecting an extensive list of research sources, I sat down to write a paper for the company that highlighted the advantages of using clickers in the classroom. Through this paper, I promoted the company's product and supported it with scientific research that showed a significant improvement in student learning outcomes.  

Readability Statistics

Passive Sentences: 4.9%

Flesch Reading Ease: 61.5

Flesch-Kincaid Grade Level: 9.3