From Zero to Certified: My Journey Through the AI Program
My background was in agricultural science. I knew how to identify tomato blight by sight, how to test soil pH, and how to fill out extension service forms. What I did not know: what a neural network was, how gradient descent worked, or what a tensor was. I enrolled in Prime Robotics AI and Machine Learning program with one question in my head: can someone like me actually learn this?
The answer is yes — but it requires honesty about where you start.
Weeks One and Two: Python Without Shortcuts
The first module does not assume you know Python. It assumes you can follow logic and want to learn. Variables, loops, functions, file I/O, error handling — all of it, from the ground up. I had written a few small scripts before, so this moved quickly for me. But I watched classmates who had never written a single line of code catch up within the first week. The instructors do not move on until everyone is comfortable. That patience matters.
Week Three: The Moment It Gets Real
Week three shifts from Python to NumPy and Pandas, then into scikit-learn for classical machine learning. Decision trees. Logistic regression. Confusion matrices. I remember feeling genuinely excited when I fitted my first model on a crop yield dataset and it predicted better than random chance. It sounds like a small thing. It was not.
By week five we were building and training simple neural networks with TensorFlow and Keras. By week seven we were doing image classification with convolutional networks. The progression is steep but deliberately staged — each new concept builds directly on the last.
The Project That Changed Everything
In the final module, you pick a real-world problem and build a solution. I chose crop disease detection from leaf photographs — a problem I understood from my agricultural background but now had the tools to approach computationally.
I collected a dataset of maize and tomato leaf images — healthy specimens, early blight, late blight, and mosaic virus. I fine-tuned a MobileNetV2 model pre-trained on ImageNet on roughly 3,000 images. Validation accuracy landed at 91.4%. Not perfect, but good enough to be genuinely useful.
The Prime Robotics instructors helped me package it into a simple Flask web app. A farmer can photograph a leaf with their phone, upload it, and get a disease prediction back in under two seconds.
After the Program
I presented the project at our cohort demo day. A representative from a Lagos-based agri-tech startup was in the audience. They contacted me the following week. Four months later the model is running in a pilot on three farms in Ogun State, and I am listed as a technical contributor.
I am not telling this story to sound exceptional. I am telling it because I was genuinely unsure whether I belonged in a technology program. If you have domain knowledge in any field and are willing to put in the work, the AI tools are learnable. Prime Robotics will teach you the tools. You bring what you already know. The combination is more powerful than either alone.
