A novel artificial intelligence model leveraging Transformer architecture, originally engineered for natural language processing, has achieved a 94.2 percent accuracy rate in identifying early-stage cardiovascular disease through electrocardiogram analysis. Published in the International Journal of Medical Engineering and Informatics, the research marks a notable convergence of machine learning and clinical diagnostics. Cardiovascular disease remains a leading global health threat, causing approximately 18 million premature deaths each year. Early detection is frequently hindered by the manual nature of electrocardiogram interpretation, which demands specialized training, consumes valuable clinical time, and carries inherent risks of human error. The developed one-dimensional Transformer model addresses these bottlenecks by processing raw electrocardiogram signals in parallel with additional patient vitals and clinical metadata. This parallel analytical framework enables the system to isolate subtle physiological patterns indicative of incipient heart disease without relying on traditional, labor-intensive waveform tracing. When paired with standard medical expertise, the model offers healthcare providers a reliable triage tool to accelerate diagnostic pathways and streamline patient routing. Researchers caution that current performance metrics are derived from established public datasets and require external validation before deployment. The next phase of development will prioritize testing on independent, demographically diverse clinical cohorts, followed by prospective trials to assess real-world robustness and regulatory compliance. Successful implementation could standardize automated cardiovascular screening across primary care and emergency settings, significantly reducing diagnostic delays and alleviating workload pressures on medical staff.