The fourth most common malignancy worldwide is cervical carcinoma (CC). HPV causes CC. In developing countries, HPV kills women. HIV-positive women account for roughly 5% of cervical cancer cases. So early detection is crucial. Traditional HPV testing and colposcopy treatment outcomes are superior. This study uses AI models to detect cervical cancer early, as they are most powerful, fast, and cost-effective. This study uses whale optimization and 1-DCNN/Bi-LSTM ensemble models to detect cervical cancer. Traditional identification processes struggle with feature optimization. Whale optimization using an ensemble deep learning model to recognize CC was our powerful weapon. By employing whale optimization, finding the best solutions and improving efficiency. The work uses VGG16, VGG19, ResNet50, and ResNet101 AI models. These popular deep learning architectures can extract important information and capture subtle patterns, improving classification accuracy. The combination of those two approaches outperforms CNN or LSTM alone, enhances accuracy, and is resilient. Experimental results demonstrate that the proposed model can handle raw, imputed, and whale-optimized gene expression data. DeepEcc 1.0 outperforms CNN designs like VGG-16, VGG-19, ResNet-50, and ResNet-101 in accuracy, sensitivity, specificity, and MCC. DeepEcc 1.0 handles high-dimensional and imbalanced datasets well with an accuracy of 85% and an MCC of 76, especially on the whale-optimized dataset. Benchmarking against recent studies using TCGA and GEO datasets shows that DeepEcc 1.0 is on par with or better than existing state-of-the-art approaches for identifying cervical cancer based on gene expression data. DeepEcc 1.0 can detect cervical cancer early and accurately thanks to Whale optimization and ensemble DL architecture, which improve feature selection and model generalization.