Dynamic Fairness in Workforce Allocation

This paper proposes a framework for fairness-aware workforce allocation by integrating dynamic feature weighting with hybrid AI models. The approach combines lexical and semantic retrieval techniques, Large Language Model (LLM)driven fairness labeling, and reinforcement learning for adaptive feature prioritization. By addressing scalability, bias amplification, and interpretability gaps in existing systems, this work bridges the divide between algorithmic precision and ethical accountability in HR decision-making. Evaluations on both synthetic benchmarks and real-world HR datasets demonstrate superior fairness-performance trade-offs compared to state-of-the-art baselines, achieving up to 91% bias reduction while maintaining 89% recommendation accuracy. Key contributions include a context-aware fairness metric, an LLM-guided reranking layer, and a stakeholder-in-the-loop weight adjustment mechanism.