Dynamic Skill Trajectory Optimization in Talent Management using a Deep Q-Network: The ATIM Framework

Escalating pace of technological change necessitates a fundamental shift from static workforce assessment to dynamic, personalized skill enhancement. This paper introduces the AI-Based Talent Improver Monitor (ATIM), an intelligent framework designed to optimize continuous employee The development. Our methodology integrates a fine-tuned BERT model for accurate skill gap identification with a Deep Q-Network (DQN) Reinforcement Learning (RL) agent—a novel application in this domain. Unlike non-adaptive systems that provide generic recommendations, the DQN agent is trained to autonomously learn the optimal sequence of actions (e.g., specific training modules, targeted mentorship) required to maximize a predefined Reward Function linked to critical organizational metrics, such as a reduction in Time-to-Competency (TTC). Empirical validation, including analysis of the agent’s rapid policy convergence (within 4,200 episodes) and a statistically significant 27.8% performance gain over conventional baseline methods, establishes the ATIM framework as a technically robust and highly effective solution for data-driven human resource management.