AI Scholars Program (Summer/Fall 2024)

As a 10th grader interested in Computer Science, especially Artificial Intelligence, I saw InspiritAI as a golden opportunity to explore AI in real-world contexts. Although I had some experience in Python and C++, I wanted to move beyond simple problems and build projects that matter. This program promised hands-on experience through socially impactful projects, which I hoped would give me a clearer understanding of AI’s potential. I was also excited to develop teamwork and communication skills while preparing for future college and academic success. 

Before the program officially began, I received setup materials to help me prepare. These included a Python crash course and an interactive notebook using the turtle library—fun and useful for brushing up on basic concepts like conditionals and loops. I also got familiar with Google Colab and some core libraries like numpy, pandas, matplotlib, seaborn, and scikit-learn. This preparation made the live sessions much easier to follow. The program itself ran over two weeks, with 10 sessions total (2.5 hours each). Each session was split into two parts: a live lecture and a hands-on notebook session in smaller cohort groups.

During the first week, we explored foundational concepts in AI and machine learning:

• Day 1: Introduction to AI and Python libraries
  We began with the big picture—what AI is and where it’s used—then practiced writing basic Python code in Google Colab.

• Day 2: Linear Regression
  This session introduced me to the ML pipeline and how a model learns from data. We worked through a notebook that visualized how predictions improve over time.

• Day 3: Classification and Logistic Regression
  I learned how to build models that classify input data into categories. It was my first time truly understanding how mathematical functions (like sigmoid) are used in AI.

• Day 4: Neural Networks and Computer Vision
  We explored how machines “see” through convolutional neural networks (CNNs). These models can recognize patterns in images better than humans in some cases. The TensorFlow Playground was especially fun to experiment with.

• Day 5: Natural Language Processing
  This session covered how AI understands and generates text. We built a basic NLP model and discussed how tools like ChatGPT are transforming communication.

There were also optional resources like video walkthroughs, quizzes, and even lectures on AI ethics and college prep—many of which I plan to revisit later.

By the end of week one, I had a solid grasp of how AI systems are trained, evaluated, and applied. I was especially excited to move into the second week, where we’d apply these skills to a real-world project.

Before the program officially began, I received setup materials to help me prepare. These included a Python crash course and an interactive notebook using the turtle library—fun and useful for brushing up on basic concepts like conditionals and loops. I also got familiar with Google Colab and some core libraries like numpy, pandas, matplotlib, seaborn, and scikit-learn. This preparation made the live sessions much easier to follow. The program itself ran over two weeks, with 10 sessions total (2.5 hours each). Each session was split After a packed first week of learning the foundations, Week 2 was all about applying what we learned. Each cohort was given a socially impactful AI project to work on, and my group’s challenge was Distracted Driver Detection—a classification problem where the goal is to determine whether a driver is focused or distracted based on input images.

We started by exploring the dataset, which included various images of drivers in different scenarios: talking on the phone, adjusting the radio, eating, or just driving attentively. Our mission was to build a model that could look at one of these images and correctly classify the driver’s status.

My main role was helping build and train the model. We used a dataset of driver photos, split it into 80% training and 20% testing, and used machine learning tools like scikit-learn to train a classification model. I worked on tuning the model to improve its accuracy, and it was really cool to see the concepts from the first week—like training, testing, and loss functions—actually in action. 

At the end of the week, each group presented their project. We created slides that explained our approach, our challenges, and our model’s performance. Presenting our work gave me a new appreciation for how important it is to communicate technical ideas clearly and effectively.

I was proud to receive a Certificate of Completion for the AI Scholars Program. More than that, I walked away with the confidence that I can actually understand, build, and explain AI systems. Week 2 didn’t just reinforce what I had learned—it proved to me that I could apply it.