Sasikiran Kaye

• Delivered lectures and created instructional materials on subjects including Fundamentals of AI, Machine Learning, Visualization Techniques, Machine Tools, and Manufacturing Engineering.
• Conducted laboratory sessions to provide students with hands-on experience and practical exposure to core concepts.
• Mentored student projects focused on the integration of Mechanical Engineering principles with Machine Learning techniques, guiding research and development efforts.

Key Projects:

Predictive Wear Analysis.

• Developed a tool-wear prediction system using force sensor data and machine learning algorithms.
• Utilized Python libraries (NumPy, Pandas), and MySQL for data preprocessing and management.
• Conducted Exploratory Data Analysis (EDA) with Tableau, Matplotlib, and Seaborn to identify critical parameters, such as load, velocity, and material composition, affecting tool wear.
• Implemented machine learning models, including KNN, Decision Trees, Bagging, and Gradient Boosting, achieving 90% prediction accuracy with a Root Mean Squared Error (RMSE) of 0.15.
• Identified Gradient Boosting Decision Tree as the most effective model for wear prediction.

Prediction of Casting Defects.

• Designed and developed an ML-based system to detect gas porosity defects in steel castings.
• Employed Python, MySQL, and Tableau for comprehensive data wrangling and visualization.
• Performed detailed EDA to discover influential factors such as pressure, speed, mold radius, and material composition.
• Applied hyperparameter-tuned models, including KNN, Bagging, and Gradient Boosting, with the final model achieving an F1 score of 0.88.
• Selected gradient boosting as the optimal approach for defect prediction.