Back in 2022, three of us decided to build an IoT-based Smart Greenhouse as our final-year project.
The idea was simple: monitor parameters such as temperature, humidity, soil moisture and light intensity, and use the data to provide appropriate feedback to the plants.
The interesting part was actually making it work.
I had always enjoyed reading about microprocessors, microcontrollers and electronics, so this project was right up my alley. The ESP32 was having its moment back then, and with built-in Wi-Fi, it seemed like the obvious choice for an IoT project.
Then came the real engineering challenge: getting all the components.
We made our way through Grant Road’s electronics market in Mumbai, hunting for sensors, wires, modules and whatever else we needed. And then there was the greenhouse itself. We somehow ended up at Navi Mumbai IKEA looking for the right “house” for our plants which was probably one of the more unusual parts of the project.
Once everything was assembled, we got to the part I enjoyed the most: writing the .ino code and watching the ESP32 interact with the sensors.
Sometimes it worked.
Sometimes it absolutely refused to work.
But getting that first proper reading from a sensor and seeing the entire system come alive was incredibly satisfying.
There was another interesting aspect to the project: this was before ChatGPT.
If the code didn’t work, there was no AI assistant to explain the error or write a quick solution. It was documentation, Google, forums, datasheets, YouTube, and a fair amount of trial and error.
Looking back, that little greenhouse project did more than get us through our final year. It sparked my curiosity about what was happening inside the boards we were using. That curiosity eventually led me to read about processors, GPUs, chips and the semiconductor industry.
Today, I spend time using GPUs to train and experiment with ML models. It’s quite amusing to think that, not too long ago, my biggest hardware adventure was getting an ESP32 to read a soil moisture sensor correctly. The scale is completely different, but the curiosity feels much the same.
From sensors and microcontrollers to GPUs and ML has been quite a journey.