Ai Resident
Current• Designed and implemented machine learning models for diverse real-world projects in NLP, computer vision, deep learning, and other AI domains• Utilized proper model tuning techniques and best practices• Benefitted from mentorship with industry-leading professionals• Built a project portfolio demonstrating end-to-end ML skills, from data preprocessing to model deployment• Example project: "Happy Customers" project, based in Python, aimed to predict customer happiness for a food delivery company using survey data. - Achieved over 60% accuracy on just 126 observations - LazyClassifier was used for initial model exploration - Hyperopt was employed for hyperparameter tuning - Tree-based algorithms like ExtraTrees and RandomForest were used - Others, like XGBoost and LGBMClassifier, were also used - Ensemble methods, like voting, achieved good accuracy, especially stacking - Key takeaways involved uncovering which questions in the survey were most predictive - I recommended that the company gathers more survey responsed for improved model performance
