Written language revolutionized human society by providing a substrate for transmission and preservation of knowledge. One of the most quintessential written formats is handwriting. Its practicality, uniqueness, and ease of use makes this format, one of the most widely used written schemes. Hence,even in the modern world, handwriting is still prevalent in a plethora of documents such as postal service labels, bank checks and old archives. Notwithstanding, in some cases, handwritten text has several pitfalls, including: illegibility, physical source damage, and difficulties transporting and editing the information. One potential way to overcome this issue is to transcribe handwritten language into digital format. Nevertheless, manually transcribing handwritten to digital is a daunting task. Here, we apply a machine-learning classification algorithm to specific features of a handwritten dataset to determine the corresponding label of a handwritten character. In order to test the efficacy of our training algorithms, we converted the input data into training and testing datasets via cross-validation. Using this approach, we systematically tested the components of our machine learning pipeline: preprocessing tools, extracted features (HOG and Hu moments) and ML algorithms (KNN and MLP). After several iterations of this process, our algorithm reliably predicted slightly above 95% accuracy the labels of the class data set.