Building financial security often begins with establishing the habit of saving. However, many Ugandans, particularly women, face the challenge of informed financial decision-making, which has caused fewer subscriptions to financial technology (fintech) application services by women than men. This disparity can be attributed to several factors, including limited financial literacy and the prevalence of English-only interfaces in many fintech applications in Uganda. Financial illiteracy hinders their ability to understand and evaluate financial products, making them less likely to adopt them. At the same time, the language barrier, which is faced by women who primarily speak Luganda or other local languages, restricts their access to critical information and instructions within the apps. Unfortunately, training and deploying machine translation models in a low-resource language setting like Uganda presents computational constraints that must be dealt with as an approach to tackle the language barrier and consequently address financial illiteracy. Thus, this research seeks to address this critical issue by developing and training context-aware, resource-efficient, and explainable transformer machine translation models tailored to the intricacies of financial literacy content in Luganda, a low-resource African language, that fintech service providers can leverage to develop more inclusive financial products and applications. To address the limited data gap, we mined data from tweets, fintech-related content on TikTok, and financial literacy-related blog posts, which we used to build a curated parallel English-Luganda corpus of over 2000 sentences focused explicitly on financial knowledge content.