The rapid diversification of technical specializations within the computer science and information technology domains presents a significant educational challenge for students, who frequently lack the profound self-awareness and practical guid-ance imperative to selecting professional pathways perfectly aligned with their inherent psychological and cognitive traits. Consequently, pivotal career deci-sions are systematically driven by external peer trends, superficial fascinations, or arbitrary assumptions rather than intrinsic behavioral suitability. This paper introduces Ascend AI, a comprehensive, artificial intelligence-driven career orien-tation framework meticulously designed to mitigate this structural uncertainty. The proposed architecture seamlessly integrates quantitative psychological pro-filing, generative LLM-based learning curriculum methodologies, and responsive, interactive audio interview simulations into a cohesive, decoupled microservice platform. Phase one of the framework processes multi-dimensional vocational preferences and personality metrics—captured via a standardized 50-item Big Five (OCEAN) inventory, a 48-item RIASEC model survey, and a distinct cognitive reading-comprehension assessment. These vectors advance through a dual-pipeline machine learning ensemble, integrating K-Means clustering and Soft-Voting Logistic Regression, to empirically predict optimal technical career trajectories. Phase two translates these discriminative mathematical predictions into highly customized, dynamically generated 10-day micro-learning roadmaps utilizing Google Gemini Large Language Models (LLMs) and strict Pydantic schema validations to ensure structured, hallucination-free knowledge acquisi-tion.