Designing encryption models for IoT deployments requires analysis of multiple network level constraints. These include, estimation of energy requirements, security strength, encryption & decryption delay, computational complexity, etc. A wide variety of models are proposed to perform these tasks, but most of them are either highly complex, or require higher energy levels for encrypting data samples. Moreover, these models are contextindependent, and cannot be used for application-specific deployments. To overcome these issues, this text proposes design of a novel secure and lightweight dynamic encryption bioinspired model for IoT networks. The proposed model initially uses an Elliptic Curve Cryptography (ECC) process for data security, and optimizes its performance via Bacterial Foraging Optimization (BFO). ECC parameters that are obtained via BFO are further fine-tuned using a Q-Learning based process, which assists in identification of context-specific parametric ranges for different network types. The combination of BFO with Q-Learning results in dynamic ECC curves, which can be used for context-specific deployments. Performance of the model was evaluated on different scaled networks, and compared with other state-of-the-art encryption models in terms of encryption delay, decryption delay, security level under different attacks, and energy consumption levels. Based on this comparison, it was observed that the proposed model showcased 8.5% lower encryption delay, 3.2% lower decryption delay, and 5.9% lower energy consumption while maintaining similar security levels. Due to these enhancements, the proposed model is useful for a wide variety of low complexity IoT deployments.