The emergence of Large Language Models (LLMs) as tools in software development processes necessitates a re-examination of foundational systems engineering practices. The systems engineering discipline has been evolving from waterfall and classic V-model life cycles to more model-based and digital engineering ecosystems including Agile methodologies and frameworks. These approaches emphasize prescriptive documentation, hierarchical decomposition and artifact-driven traceability from requirements to implementation and testing. While effective in human-centered engineering workflows, these methods exhibit limitations when development tasks are increasingly executed by humans supplemented with LLMs, and, prospectively, by LLMs alone. LLMs require well-structured instructions and contextual narratives rather than visual diagrams, lengthy spreadsheets, or specification documents. Although good prompt engineering focuses on crafting quality input text based on syntax, phrasing, and structured input, prompts are limited by the model’s training and context. To improve the output from LLMs for the purpose of software engineering, this paper advances the concept of Context Engineering, defined as the systematic curation, organization, and management of contextual information to guide Artificial Intelligence (AI)-augmented software development. Narrative-rich artifacts like user stories and scenarios along with LLM-focused interaction patterns, knowledge bases, and embeddings supersede rigid requirements documents as the primary methods for conveying intent and information. With better context, LLMs can better realize the intended software systems. The paper begins with foundations of systems engineering and LLMs to describe the current disconnects between systems engineers and generative software development. Context Engineering is then explained and proposed as an evolutionary trajectory of systems engineering. The paper will include some implications for the engineering processes that need to evolve in an era where development is increasingly delegated to LLMs. This paper is not an in-depth examination of LLM technology, but rather how systems engineering can evolve using LLM technology.