During my internship at Beehive Drones, I worked on the data processing team, focusing on analyzing aerial imagery for environmental applications. Specifically, I contributed to a project aimed at estimating carbon values from drone-captured images.
Key Contributions
- Data Pipeline Development: Developed a robust data pipeline to process raw aerial imagery. This involved preprocessing, cleaning, and extracting relevant features necessary for carbon estimation models.
- Model Engineering: Engineered and trained a deep learning model to predict carbon values. Through iterative experimentation and optimization, I achieved a 93% improvement in model performance over the initial baseline.
- Collaboration: Collaborated with the drone operations and software teams to ensure the data collection process aligned with the modeling requirements.
Learnings
This internship provided me with hands-on experience in handling real-world unstructured data (images) and applying deep learning techniques to solve environmental challenges. It sharpened my skills in Python, TensorFlow, and data engineering best practices.