Hardik Ravindra Sakpal

Inspired by the widespread application of machine learning across engineering, medical, and computational domains, motivation has grown to become an AI enthusiast with a passion for solving real-world applications. Adept in ML libraries, demonstrated skills in AI techniques and Big Data Analysis have been shown. Consistently seeking opportunities for continuous learning and collaborative research in cutting-edge technologies. The goal is to make significant contributions in the field of ML and help advancements in AI.

Student Project (Studienarbeit), Institute of Particle Technology, TU Braunschweig, 

Development of an AI-based Process translator framework to optimize the Efficiency of Battery Materials Production, 

• Trained neural networks on data from distinct machine setups and developed a Translator model through ensemble methods, achieving 98% accuracy by combining insights from two datasets for feature extraction on similar new data types.

Machine Learning Industry-Project, Deutsches Zentrum für Luft - und Raumfahrt, Braunschweig, 

• Neural networks for leakage detection in vacuum bag processes, Achieved over 87% accuracy in leakage detection by training and comparing equivariant versus standard network architectures, with data augmentation improving robustness across dataset sizes; consistently secured 90%+ results via hyper-parameter optimization and architecture selection.
• Multiclass classification EUROSAT, Conducted multiclass classification on the EUROSAT dataset using Python, pre-processing images i.e. resizing and normalizing for training/testing, and developed Neural Networks for evaluation