As: ML Engineer
- Built and trained CNN models in TensorFlow for classifying plant diseases with 94%–98% accuracy.
- Processed 1,600–3,200 training images and 400–800 validation images per model across 100 epochs.
- Used Adam optimizer (learning rate: 0.001) for efficient training and optimal performance.
- Conducted data preprocessing and visualization with NumPy, Matplotlib, and Seaborn.
- Managed datasets via Google Drive, using Jupyter Notebook and Google Colab for experimentation.
- Tools: Python, TensorFlow, Keras, NumPy, Matplotlib, Scikit-learn, Google Colab.