From student project to start-up: RP grads develop AI-powered camera system for stroke detection

SINGAPORE – In 2005, when Ashwin Manikandan was born, his grandmother suffered a stroke, but his family did not realise her facial features were drooping. 

By the time they rushed her to the hospital, the damage was severe enough to leave her permanently disabled. 

Describing the deterioration of his grandmother’s condition in the years after her stroke, Manikandan, now 21, said: “She started to have pain (throughout) her body, then had difficulty walking. Slowly, it became harder to move, until she finally became bedridden.”

“I don’t want that to happen to anyone else,” said the Republic Polytechnic (RP) graduate about his grandmother, who died in 2021. “I had to grow up seeing her always suffering from pain and unable to move at all. There was nothing we could do except make her feel as comfortable as possible.”

He cited the experience as one of the driving forces behind an AI-powered stroke detection system developed by him and three other schoolmates: Tan Jia Ling Jolin, 20, Sasindren Jaya Sanger and Sarah Beegam Saju, both 21.

In 2025, they met at an entrepreneurship bootcamp and won the pitching competition. They decided to continue pursuing the stroke-detection idea in RP’s Entrepreneurial Immersion Programme (EIP) and received $3,000 in funding to come up with a minimum viable product.

The project, called Stroke Guard, utilises an AI-powered camera to detect facial drooping, one of the key signs of stroke. This aims to reduce the need for round-the-clock patient monitoring and to supplement caregiver assessment of stroke signs. Once the AI has detected potential facial drooping, it sends an alert to the patient’s caregiver to conduct further assessments.

At its current stage, Stroke Guard is for bedridden patients in nursing homes. Cameras are installed at fixed positions pointing towards the bed, so the detection system can accurately determine facial drooping.

To develop their product, the team collaborated with charity organisation Care Corner’s Active Ageing Centres. From the centres’ clients, they collected more images of facial expressions to train their AI and reduce false positives. The team collected 70 datasets from this collaboration.

“They were attracted to the idea because it would be very helpful (for the elderly),” said Tan, adding that Care Corner has also expressed interest in the completed Stroke Guard product. 

In total, it took the team about four months to develop a working prototype under the EIP.

In May, they attended the Danang Venture and Angel Summit (DAVAS) to meet potential angel investors and learn more about the medical technology market in Vietnam.

After their pitch at DAVAS, an investor offered to connect the team with Vietnamese hospitals. “We were also invited to collaborate with another company to form an international start-up,” said Tan.

Since graduating in May, they are continuing to develop Stroke Guard as a start up with RP’s Office of Entrepreneurship Development. The team plans to commercialise the product in the future.

Here are three takeaways from the team you can apply to developing your own product.

1. Dig deep into your experiences to solve a problem you and others face

Other than Manikandan’s personal experience, Tan – the main mind behind the Stroke Guard concept – also experienced the intense and time-consuming nature of caregiving firsthand.

In 2023, her younger brother, then aged 10, was diagnosed with febrile infection-related epilepsy syndrome (FIRES), a syndrome that causes seizures to develop following an illness marked or caused by fever.

It is a rare condition and the seizures are usually resistant to treatment. Tan’s brother was admitted to the ICU in 2023 due to the severity of his condition.

Tan took care of her brother full-time for four months after her O Levels. Once she started her studies at RP, she realised it would be difficult to juggle both caregiving and school. For a while, she thought she would have to sacrifice her studies.

“The situation changed after school started. (My father) wanted me to focus on my studies, and didn’t want me to sacrifice them. So, he stopped working to become my brother’s full-time caretaker,” said Tan.

This inspired her to think of solutions that would alleviate the burden on caregivers for patients who required similar round-the-clock care.

“Technology can help us do a lot of things that humans can’t,” she said. “It can help caregivers get more rest and be more efficient.”

2. Find at least one point of difference from what is on the market

“A lot of (medical technology) right now consists of wearables, or has to physically touch (the patient),” said Sarah.

“Stroke Guard is just a camera system in the corner of a room or at a patient’s bedside, so it doesn’t disrupt patients’ rest or sense of movement.”

3. Be prepared to learn along the way and get hands-on

As engineering students, the Stroke Guard team had no experience in coding or app-building.

Technical lead Sasindren turned to YouTube videos to learn Python and HTML in order to code their software.

Tan and Manikandan worked together to refine Stroke Guard’s AI detection by feeding in the collected datasets. 

“Initially, we thought the technical part would be the hardest,” said Tan. “But the biggest hurdle we faced was actually the business part.”

Sarah, the designer, handled most of the business-related aspects. With the help of the Internet, she came up with a business model for their start-up.

She said: “I had to come up with a profit model: Would we be a subscription- or product- based business? What would our customer and supplier segments be?”

The team’s technical mentor always challenged them to set bigger goals and try more daunting things. Tan recalled: “He would always ask us: What’s stopping you from doing this? Have you tried? You’d be surprised by how much people can help if you just ask.”

For example, the team’s collaboration with Care Corner’s Active Ageing Centres came about only because they reached out to the organisation first.

Correction note: This story has been updated to correct the programme name to Entrepreneurial Immersion Programme.