Many engineering students struggle to produce coherent, well-structured scientific papers even when they understand technical content. This study investigates how a five day instructional bootcamp that integrates Generative Artificial Intelligence (GenAI) as a pedagogical scaffold can improve students' scientific writing outcomes. The intervention guided students through the complete research paper structure - introduction, methods, results, and discussion - while using GenAI tools to model organization, suggest revisions, and prompt meta-cognitive reflection on clarity, coherence, and evidence use. A mixed methods analysis combined pre and post survey data with qualitative reflections. Quantitative results show significant gains in students' ability to plan and outline scientific texts, construct logical arguments, revise for accuracy, and integrate credible sources. Qualitative findings confirm higher confidence in writing, improved understanding of academic conventions, and stronger ethical awareness of authorship. Students described GenAI as a reflective partner that helped transform writing from a mechanical task into an iterative reasoning process. These outcomes indicate that GenAI, when framed as a scaffold rather than a shortcut, can effectively cultivate scientific writing literacy, structural competence, and confidence among engineering undergraduates.
