This chapter explores the AI alignment problem at the intersection of machine learning and philosophy, aiming to clarify a concept often clouded by abstract speculation and philosophical overreach. While alignment is widely acknowledged as central to AI and ML safety, its interpretation is frequently hindered by associations with intractable metaethical debates and speculative AGI narratives. Hence, I propose reframing alignment, presenting it as a concrete technical limitation, grounded in current gradient-based learning paradigms, rather than an abstract philosophical puzzle. Simultaneously, we show how alignment necessarily involves the epistemic challenge of learning and structuring human preferences. By integrating technical and normative perspectives, we will also learn to distinguish between outer and inner alignment and argue that progress demands philosophical reasoning and ML methodology engagement. This chapter provides foundational knowledge for newcomers, stripping away unfalsifiable assumptions and focusing on alignment as a real and addressable problem.