At Codtech, I built predictive models in Python for classification and regression tasks, and designed data pipelines for preprocessing and analysis. Model tuning improved accuracy by up to 8%.
For my Forest Fire Risk Prediction project, I developed a Random Forest model using meteorological, satellite, and historical data. It reached 92% accuracy, and I evaluated it with accuracy, precision, and recall.
I also developed an SRGAN image restoration project using TensorFlow, Keras, and OpenCV, applying PSNR and SSIM to evaluate image quality. My work included preprocessing pipelines and tuning the generator–discriminator balance and loss functions.
