Things I Wish I Had Known Before Starting ML

Pascal Janetzky reflects on his experiences in machine learning, highlighting key aspects often overlooked. A significant portion of ML work involves preparing data, which includes cleaning, aligning, and debugging datasets, rather than just model training. Writing effective research papers is akin to crafting a sales pitch, requiring a tailored approach to resonate with the target audience. Progress often involves debugging existing code and adapting it to new environments, rather than creating everything from scratch. While breakthrough discoveries are rare, consistent effort and acceptance of incremental progress are essential for a fulfilling career in machine learning.