We present SkiTechCoach, a novel alpine skiing dataset designed to advance data-driven skill assessment and coaching. Our dataset includes sole pressure and 3D pose measurements from an indoor ski simulator, along with two objective performance indicators: FIS points and coaching commentary. The dataset comprises 160K video frames of 71 high-quality sequences of 20 expert skiers performing slalom turns at varying speeds, synchronized with sole pressure, pose trajectories, FIS points, and coaching feedback delivered via verbal and visual annotations. To demonstrate its utility, we establish baselines for FIS point estimation using both pressure and pose modalities, and we explore automated coaching generation through multimodal large language models. Our empirical analysis reveals that existing LLM-based coaches lack the domain expertise required for nuanced alpine-ski feedback, underscoring the need for specialized model fine-tuning. We release SkiTechCoach to catalyze future research on ski skill acquisition and intelligent coaching systems that leverage synchronized wearable pressure sensing and markerless motion capture.