This 14-Year-Old Built an App That Detects Heart Diseases in Seconds
Siddarth Nandyala wants to put his tool in the hands of medical professionals so that they can catch cardiovascular abnormalities in their early stages
/https://tf-cmsv2-smithsonianmag-media.s3.amazonaws.com/accounts/headshot/Ramsha_Zubairi-edit.png)
In trials in India, Siddarth Nandyala detected and diagnosed more than 40 patients with potential cardiovascular diseases, each within a span of seven seconds. Not a doctor, or a medical student for that matter, the 14-year-old was able to do this thanks to his very own invention, a simple smartphone-based app called Circadian AI.
Last year, for nearly eight months, the Texas-based teenager spent hours huddled over his computer. He had one simple goal—invent an application that can pinpoint cardiovascular abnormalities during their initial stage.
In most cases, a heart attack or stroke is the first sign of a cardiovascular disease. Initial-stage cardiovascular abnormalities are often asymptomatic, and early detection generally depends on routine health checkups, electrocardiograms, stress tests, echocardiograms and blood tests. Certain diagnostic tools like cardiac MRI with late gadolinium enhancement, invasive coronary angiography, right heart catheterization and biomarkers of end-stage damage are only able to detect cardiovascular abnormalities in cases where the patient is already in the late stage and suffering from severe blockages, heart failure or dead tissues.
“The main focus and goal for me out of this was to essentially create a tool that is able to help a large amount of people just through non-invasive screening procedures,” says the rising freshman at University of Texas at Dallas. He’ll be pursuing a bachelor’s degree in computer science.
Since his early childhood years, Nandyala has been interested in technology, coding and engineering. In 2022, the student designed a prosthetic arm that costs merely $150 in India, as opposed to the traditional ones that can cost up to $30,000. A year later, he founded STEM IT, a startup that creates and sells science, technology, engineering and math kits to students. The local Frisco Chamber of Commerce named him Innovator of the Year and awarded the then-seventh grader more than $2,000 to support his company. He also received a Certificate of Recognition from the U.S. House of Representatives and a letter of congratulations from then-President Joe Biden. Then, in 2024, he invented an armband that detects falls in elderly citizens with a 96.1 percent accuracy rate, higher than the Apple Watch.
“What really took my interest in the health care side of artificial intelligence was the sheer amount of impact and the change that can be made,” he says.
Heart disease is the leading cause of death globally, according to the World Health Organization. Cardiovascular disease caused 19.91 million deaths in 2021, which is almost one in every three deaths globally. A new study published in the Journal of Cardiac Failure revealed nearly 6.7 million Americans aged 20 and above suffer from heart failure. This number is expected to grow to 8.7 million by 2030, 10.3 million by 2040 and 11.4 million by 2050.
“Even one life detected is one life saved,” Nandyala says.
Early detection of disease can often slow or reverse its damage. An early intervention creates the possibility of improving a patient’s quality of life, reducing mortality rates and the cost of health care. It effectively eases the burden of the disease on the individual. However, in the absence of primary health care programs in many low- and middle-income countries, early detection and, at times, proper medical treatment is not always a possibility. This lack of access to effective and equitable health care services along with other socioeconomic factors ultimately results in an increased mortality rate in such countries. Studies have revealed that 80 percent of global deaths caused by cardiovascular diseases occur in low- and middle-income nations.
/https://tf-cmsv2-smithsonianmag-media.s3.amazonaws.com/filer_public/54/fd/54fd800e-079c-4207-bf16-5d361212d452/img-20250530-wa0001.jpg)
To develop a real working product, Nandyala frequented various hospitals and institutions, discussing issues related to cardiovascular disease with hospital staff, patients and others. There, he collected the necessary data for his app. The next step was to create a robust artificial intelligence model, training and then retraining it again until he was completely satisfied with the results.
“I’d definitely say that throughout this journey every interaction was helpful,” Nandyala says.
The app is currently only available for clinical use by authorized personnel. It’s just like any other medical device, Nandyala explains. “You need to know how to use it before you can use it to detect any diseases,” he says.
To operate the app, the user simply places a smartphone near their heart, where it records the sound of the heartbeat. The app has enhanced noise-cancellation techniques to filter out ambient sounds, enabling accurate readings even in loud environments. The recorded sounds go through various amplification algorithms and are then sent over to a machine-learning model hosted on the cloud, which gives an overall synopsis of the user’s heart health. An intuitive interface displays the results, with explanations about what is normal and what may need medical attention. The app detects arrhythmias and irregular heartbeats, early signs of heart failure, indicators of coronary artery disease and heart valve abnormalities.
“It’s a pre-screening tool,” says Nandyala.
The teenager has tested the app at various points of care in both the United States and India. In the U.S., the testing was done on around 15,000 patients, while in India the number sits close to 3,500.
“Clinical trials were conducted not only for detecting diseases, but as a metric for the efficacy of the tool itself,” says Nandyala.
During one such initial test conducted in India at Government General Hospital in Guntur, the app detected and diagnosed around ten patients with cardiovascular diseases. The diseases were later confirmed via clinical validations, through electrocardiogram and 2D echocardiograms.
The app reportedly has a high success rate in detecting heart abnormalities. “Obviously, there is diversity in the results from test to test, but we did get over 96 percent accuracy,” says Nandyala.
Currently, Nandyala is in the process of acquiring various regulatory approvals. From there, Circadian AI plans to undergo a full-scale rollout throughout the U.S., India and other countries at various primary health care centers and in clinical environments for mass screenings.
“There is incredible potential for heart-monitoring devices and apps such as Circadian AI to advance medical care,” says Jameel Ahmed, an electrophysiologist at Louisiana State University in New Orleans. Nandyala’s app can reach populations where access to care is limited, he adds. “It is particularly impressive that a 14-year-old student can create such an application that has the potential to affect millions,” he says. “With advancement in A.I. and improvements in the technology readily available to millions, we will only see more such devices and programs.”
The device, Ahmed says, is attempting to diagnose cardiac conditions based on potential irregularities in detected heart sounds. “An early potential diagnosis in patients who otherwise may not have had access to medical care may ultimately reduce long-term morbidity and mortality from this condition,” he says.
Yet, Ahmed also considers the app’s shortcomings. One potential limitation may arise from the “quality and fidelity” of the microphone on the device being used, he says. Historically, he explains, medical professionals have used stethoscopes to listen for particular sounds and frequencies of sounds at specific times within the cardiac cycles to make diagnoses. “Sometimes, these patterns, even with an expert, can be inaccurate or non-specific, and we would rely on echocardiography, imaging of the heart, to confirm,” he says.
Nandyala clarifies that the app only detects cardiovascular abnormalities and is not meant to detect what type of cardiovascular disease a person is suffering from. “We are not saying that we will replace any type of machinery, like [electrocardiograms] or 2D echo,” he says. It is simply a pre-screening tool that gives an overall analysis of whether users have any issues in their heart.
The young innovator is already expanding on the device. He’s currently working on applying it to the sounds of lungs, to detect conditions such as pulmonary embolism, water retention and pneumonia.
“I don’t want to stop here,” he says.
