One team is preparing for a cybersecurity threat that may become serious only when quantum computers become powerful enough to break current encryption. Another is trying to find para-athletes in smaller towns who may never enter formal scouting systems. In Kerala, a student is building a network to give dam operators a clearer picture of what is happening across a river.
The technologies are different. So are the problems.
What connects them is that none of these challenges is particularly new or obvious. They sit in the gaps between existing systems. Some are too early to attract attention. Others involve communities that are often overlooked. Some simply need better ways to collect and connect information.
These projects are among 40 teams selected as semi-finalists for the fifth edition of Samsung Solve for Tomorrow in India. The teams come from 15 states and 26 cities, with more than half from Tier 2 and Tier 3 locations. Twenty-five of the 40 teams have already started building prototypes.
The cohort spans AI, healthcare, education, environmental sustainability and sports. But the more interesting pattern is not how many teams are using AI. It is what they are choosing to use it for.
Preparing For A Threat That Has Not Arrived
Cybersecurity is one of the less visible problems being tackled by the cohort.
QuantHunt, a team from Raebareli, Uttar Pradesh, which participated at the event, is working on the threat known as “Harvest Now, Decrypt Later” (HNDL). The idea is relatively simple. Attackers can steal encrypted information today and store it until quantum computers become capable of breaking the encryption.
That creates a problem even before large-scale quantum computing becomes practical.
“As quantum computing advances, sensitive financial, medical, and personal records stolen today could be decrypted later through ‘Harvest Now, Decrypt Later’ attacks,” said team members Akul Attre and Saksham Shreyans.
QuantHunt is building a tool that uses AI to analyse codebases and host domains for weaknesses in encryption. The aim is to identify vulnerabilities before they become exploitable by future quantum systems.
The challenge for the team is not only technical. Any wider adoption would require integration with existing developer tools and alignment with security standards and regulations.
The project is therefore looking at a problem that most businesses may not yet see as urgent: how to protect information today against computing capabilities that may mature years from now.
Finding Athletes Outside The System
Another team, ParaSync, is tackling a more immediate gap.
Built by Code Blooded from Sundargarh, Odisha, the project uses computer vision to identify para-sports talent in smaller cities and rural areas. Athletes in these locations may have limited access to selection trials, coaches, sports science facilities and medical support.
The team wants to turn a smartphone into an initial screening tool.
“ParaSync aims to turn a basic smartphone into an accessible screening and talent-identification tool,” said team members Vatala Harika and Ritesh Routray.
The system uses computer vision trained to account for non-standard body mechanics and wheelchair dynamics. It is designed to assess factors such as propulsion posture, symmetry and fatigue.
They are working towards a lightweight version that can function offline or in low-bandwidth environments. But the bigger problem is access. ParaSync is trying to make the first stage of identification more accessible. The next step for the team is to explore partnerships with para-sports federations and grassroots academies.
Not Every Problem Needs AI
The cohort also offers a useful counterpoint to the current rush to put AI into almost every technology problem.
AcoustiNode Systems from Thrissur, Kerala, is working on a problem that predates AI.
A dam operator may know the water level at a reservoir without having a complete picture of how much water is arriving upstream or moving downstream. Without that information, it can be harder to understand how conditions are changing across a river system.
Student innovator Novain Shibu built a system using three connected nodes to monitor upstream inflow, dam levels and downstream flow. The information is brought together at a central hub and transmitted over licensed VHF radio.
That means the system does not depend on mobile networks.
“AcoustiNode Systems closes that gap with three connected nodes — upstream inflow, dam-level, and downstream flow — correlated together at the dam hub into one combined picture,” Shibu said.
The system has already been field-tested at Kerala’s Peechi Dam. The ambition is to connect systems across parallel river networks and eventually move towards basin-scale flood management.
There is no need for a large language model here. The innovation lies in collecting fragmented information and making it available in one place.
That distinction is important. As AI becomes the default answer to many technology problems, some of the most useful solutions may simply involve better sensors, communication networks and data.
Building Outside The Technology Hubs
The projects also reflect where some of this experimentation is happening.
The 40 semi-finalists come from cities including Varanasi, Amritsar, Rajkot, Guntur, Thrissur and Coimbatore. More than half are from Tier 2 and Tier 3 locations.
Their problems are often shaped by their surroundings.
For a team in Sundargarh, the question may be how to identify sporting talent that does not reach national selection systems. In Kerala, it may be how to monitor a river when mobile connectivity cannot be assumed. In Raebareli, it may be how to prepare organisations for a security threat that has not yet become mainstream.
The geography does not guarantee that these solutions will work. But it can influence which problems innovators notice first.
Other teams are working on accessible education, healthcare and agriculture.
AIRMED from Guntur, Andhra Pradesh, is exploring an AI platform intended to connect overseas patients with Indian hospitals and support India’s medical tourism ecosystem. Other projects include assistive technology for visually impaired swimmers, an AI companion for visually impaired students, interactive learning systems that use existing classroom projectors, and technology for converting agricultural biomass into clean energy and carbon resources.
Starting With The Problem
The teams are now moving from ideas towards prototypes.
Dr Nikhil Agrawal, Managing Director of the Foundation for Innovation and Technology Transfer at IIT Delhi, said 25 of the 40 teams have already started building prototypes.
“The transition from an interesting idea to a viable solution is often the most difficult part of the innovation journey,” Agrawal said.
The next stage will require teams to test their assumptions, learn from users and determine whether their technology can work outside a controlled demonstration.
That is often where technology projects become real products, or stop being projects at all.
The approach is simple: understand the problem first, then decide whether AI is actually the right tool.
The 40 teams will enter an Innovation Bootcamp in September. They will receive prototyping and mentoring support before the National Pitch Event narrows the cohort to 20 teams. The eventual winners will receive funding, hardware and incubation support at IIT Delhi.
But the more interesting measure of this cohort will come later.
Most of these ideas are still early. A benchmark, prototype or demonstration does not establish that a solution will work at scale. ParaSync will need to show that its computer vision system can identify real sporting talent. QuantHunt will have to demonstrate that its security tool can find meaningful vulnerabilities. AcoustiNode will need to prove that its network can operate reliably across larger river systems.
Their technologies may change as they build. The problems are likely to remain. And that may be the most useful lesson from this group of young innovators. At a time when AI is often treated as the starting point for a new product, some of India's youngest builders are beginning somewhere else.
They are starting with problems that are easy to overlook. Some are looking years ahead. Others are looking at people who are routinely missed by existing systems. And some are asking whether an old infrastructure problem can simply be measured and connected better.
These are not the problems that usually dominate India’s technology conversation. But for the young builders working on them, they are problems worth solving.
Your reaction