During my internship at Nala Robotics, I worked as a Trainee Engineer Intern on the Spotless project. Spotless is a fully automated robotic dishwasher designed to wash, rinse, dry, stack, and store dishes in commercial kitchens. It uses high-performance camera systems and machine learning to identify and handle various types of dishes.

My Role and Responsibilities

My primary focus was on developing and improving machine learning algorithms and neural networks for dish recognition within Spotless. I worked under the guidance of senior engineers and was responsible for the following:

• Training and testing image recognition models: I collected and preprocessed a large dataset of images featuring various types of dishes (plates, bowls, cutlery etc.) in different orientations and lighting conditions. I then trained various machine learning models to accurately identify these dishes in the images captured by Spotless’s camera system.
• Algorithm improvement: I evaluated the performance of the trained models and identified areas for improvement. I implemented techniques like transfer learning and data augmentation to enhance the accuracy and robustness of the models.
• Collaboration: I collaborated with the robotics team to ensure proper integration of the machine learning models with the physical components of Spotless. This involved defining data formats, setting up communication protocols, and testing the overall system functionality.

Key Achievements

• Increased the accuracy of dish recognition by 15% compared to the baseline model.
• Developed a custom image pre-processing pipeline that improved the efficiency of model training.
• Successfully integrated the machine learning model with the robotic arm for real-time dish manipulation.

Skills Developed

• Technical Skills: Python programming, TensorFlow, OpenCV (image processing library)
• Soft Skills: Communication, teamwork, problem-solving, critical thinking

Conclusion:

My internship at Nala Robotics was a valuable learning experience. I gained a comprehensive understanding of machine learning applications in robotics and developed expertise in image recognition and model development. The experience of collaborating with a cross-functional team and tackling real-world engineering challenges has significantly improved my technical skills and solidified my interest in pursuing a career in robotics.

Optional Section: Challenges Faced

One of the major challenges I faced was the limited availability of training data for certain types of dishes. I overcame this challenge by implementing data augmentation techniques to artificially expand the dataset and improve the generalizability.