How My Grandfather’s loss of Sight Led Me to a Research Breakthrough

A couple of years ago, my grandfather began losing his vision. At first, I didn’t know how I could help — I was just a high schooler, and this felt like something far beyond me. But then I found a way to channel that feeling into something real.

Six months ago, I applied and was selected for a research mentorship at the University of Cambridge, where I worked with Dr. Abdulali on a project that changed how I see engineering and what I want to do with it.

The Research Project

The title of my paper is:“Improving Visually Impaired Safety: Predicting Guide Dog Distractions with Pose Estimation and Machine Learning.”

It sounds intense, because it was. Over twenty weeks, I designed, built, and tested a machine learning model that could detect when a guide dog was getting distracted, in real-time, using video footage and body pose estimation.

Solving a Real Problem

The project stretched me beyond anything I’d done before. I started by interviewing expert trainers at Guide Dogs for the Blind to understand how distractions happen and what they look like. Then I got to work:

• Pose estimation: Mapping the coordinates of moving dogs’ body parts from real video footage

• Event detection: Defining what a “distraction” even is, mathematically, using anomaly detection

• LSTM modeling: Training a Long Short-Term Memory neural network to recognize patterns and predict those distractions in milliseconds

I evaluated my model using F1, precision, and recall scores,and the results showed strong predictive accuracy. What started as a deeply personal problem turned into something real, technical, and impactful.

Recognition and What’s Next

My paper is currently being submitted to a peer-reviewed accessibility journal, and I was honored with a STEM Research Award from the Cambridge Center for International Research for promising engineering work.

More importantly, this experience confirmed two things for me:

1. I love solving open-ended, real-world engineering problems

2. I’m driven by challenges that matter especially ones tied to accessibility, energy, or social good

This wasn’t just a school project, it was an introduction to what it means to do research that doesn’t yet have an answer. And that’s exactly what I want to keep doing.