The global shift from defined benefit to defined contribution pension plans has emphasized the importance of individual retirement saving decisions. In response, India introduced the NPS All India model to address the retirement needs of its unorganized labour force, offering subscribers a choice between the Auto Choice (life-cycle approach) and the Active Choice (customized allocation). Recognizing the financial literacy gap among subscribers, this study employs a Genetic Algorithm (GA) to determine the optimal asset allocation weights for Active Choice. The results demonstrate that the GA-optimized portfolio outperforms the Auto Choice in terms of expected returns, accumulated retirement wealth, and monthly pension, while maintaining manageable risk levels. These findings have important implications for policymakers and fund managers in enhancing retirement outcomes for unorganized sector workers.