After His Parents’ Motorcycle Accident, 15-YO Built an AI Tool to Fix Delhi’s Potholes

Every time we navigate around a pothole on India’s roads, swerving our two-wheelers at the last second or bracing ourselves for that familiar jolt in our cars, we go through the same mental routine. We curse under our breath, wonder why the roads are so terrible, ask ourselves why nobody fixes these things, and then resign ourselves to the reality that nothing will change.

But for 15-year-old Parth from Delhi, that familiar cycle of frustration took a different turn. When potholes stopped being an abstract annoyance and became something that directly affected his family, the helplessness transformed into urgency, and the complaints into action.

It was supposed to be a happy occasion. Parth’s parents had travelled to Agra to celebrate his grandfather’s anniversary, spending the day surrounded by family and memories. But their journey home that evening turned what should have been a joyful day into something far more frightening.

Riding back on their motorcycle through the darkness, they struck an unfinished construction site jutting out from the road. His father lost control, and they fell.

“Thankfully, it wasn’t a major accident, but it did awaken a calling inside me that I really needed to do something about this,” Parth recalls.

That internal fire would eventually become Project Sadak, a citizen-driven platform that is accomplishing what seemed impossible: actually getting potholes repaired, and doing it faster than the government systems designed for exactly that purpose.

When personal experience meets technical skill

Parth isn’t spending his Class 11 year at Saint Francis School in Delhi the way most teenagers do. While managing physics, chemistry, mathematics and computer science coursework, he’s also preparing to present a research paper at the Society for Study of Artificial Intelligence in the UK this July.

His academic interests lean heavily towards artificial intelligence and computer science, the kind of abstract, future-focused work that impresses university admissions committees.

But it was something far more immediate that redirected those technical skills towards solving a problem he could see, touch and witness firsthand. Once he started researching the scope of India’s pothole problem, retreat became impossible.

Approximately 20,000 people die annually in India from pothole-related accidents. These weren’t just statistics. They represented real people, real families and real lives disrupted by infrastructure failures that everyone accepts as inevitable.

He did what any frustrated citizen might do first: he checked what systems already existed. The Municipal Corporation of Delhi has an app where people can report potholes. Parth downloaded it, tested it, filed reports and waited.

“It takes a lot of time. In many cases, it takes three to four months before you get a reply,” he discovered. “By the time authorities respond to a pothole report, the monsoon may have made it worse, or more accidents may already have occurred.”

The gap between what existed and what was needed became impossible to ignore. Government systems weren’t failing because the technology didn’t exist. They were failing because of a lack of urgency, absent accountability and broken follow-through.

In January 2026, Parth sat down at his computer and started building what he wished had existed.

Building a platform from scratch

Project Sadak began with Parth writing code alone, line by line, teaching himself what he didn’t already know. By mid-February, it was ready to launch.

“The platform was built by me from scratch,” he explains. “Initially, yes, I started writing code, but as the complexity began to rise, I had to make use of AI.”

The platform that emerged combines artificial intelligence, citizen reporting and direct government outreach in ways that existing systems somehow miss despite having far more resources.

The core innovation lies in making sure reports are both genuine and actionable. When someone submits a pothole report, they upload a photograph. An AI system immediately analyses whether the image actually shows a pothole, filtering out spam, pranks and mistaken reports before they clog the system.

If the AI confirms it is legitimate, the report goes live immediately. If the algorithm is uncertain, it flags the submission for human review.

It’s the kind of elegant solution that seems obvious only after someone builds it, addressing the quality-control problem that plagues most crowdsourced platforms.

Users classify the potholes they report into three severity levels: high, medium and low. But Parth didn’t trust self-reporting alone. He and his small team review each classification, overriding it when necessary.

“The user may have marked it as high severity, but based on our assessment and the approach we’ve been following, we believe it should be categorised as medium severity, so we can change it,” he explains.

This human oversight ensures that truly dangerous potholes don’t get lost in a sea of minor complaints and that emergency repairs are prioritised appropriately.

A lean team with clear roles

While Parth handles the technical development and overall management, keeping the platform running and evolving, he’s built a compact team around him. One friend manages quality control, serving as the human checkpoint for every report that comes through the platform.

Another scouts for grant opportunities, helping keep the operation funded without compromising its independence. Three people in total, all students, managing what would normally require a full organisation.

The metrics reveal a platform gaining genuine traction: 307 unique open reports, between 250 and 500 daily visitors, and a geographic reach that now spans Delhi, Bengaluru, Gurugram and other parts of Haryana.

But numbers tell only half the story. What matters more is what happens after someone hits the submit button.

From report to repair: Making bureaucracy actually work

Once a pothole report clears AI detection and administrative review, Project Sadak’s automated system takes over in ways that government apps somehow never manage.

The platform generates detailed emails and sends them directly to the relevant authorities, including the Municipal Corporation of Delhi and the Development Department of Delhi, depending on which jurisdiction the pothole falls under.

“Initially, we were emailing the parties manually,” Parth admits, “but now the platform automatically writes an email and submits the report.”

This automation solves a problem that defeats most citizen activism: the exhausting burden of follow-up. People report problems once, maybe twice, but few have the time or energy to send repeated emails, make phone calls, document non-responses and escalate issues through bureaucratic channels.

Project Sadak does this automatically, creating a paper trail for every single report and making it harder for authorities to ignore problems through simple inaction.

But Parth discovered something that every civic activist eventually learns: sometimes official channels aren’t enough, and sometimes you have to become the solution yourself.

Of the 11 potholes fixed through Project Sadak so far, Parth personally managed the repair of 10 in Delhi.

He used grant money to hire two to three local workers, purchased the necessary materials and showed up to oversee the repairs himself. His father’s construction industry connections made hiring labour straightforward, but it was Parth’s determination that turned reports into actual filled potholes.

“I would reach there myself to repair them, and I have pictures with almost all of the potholes,” he says excitedly, documenting his work the same way he documents every report on the platform.

The eleventh repair happened through what Parth calls magic, though viral attention might be a more accurate description.

“In Bengaluru, a reported pothole got fixed even though the official channels seemed unclear. We couldn’t reach the authorities, but even without that, they fixed it on their own,” he says.

Parth credits this partly to Mukund Jha, an entrepreneur who reposted one of his posts about Project Sadak, suddenly bringing visibility to the platform and flooding it with reports from Bengaluru.

Sometimes change happens through official channels, sometimes through public pressure, and sometimes through a combination nobody can quite explain.

Building partnerships that could scale impact

Parth’s vision extends far beyond personally filling potholes with hired workers.

“My future plans are to expand this to other Indian states, Asian countries and African countries because these are the ones that are majorly affected by potholes,” he explains.

“If I happen to raise enough grants, I’ll try my best to maybe repair these potholes myself and, if possible, construct new roads for people.”

It’s the kind of ambitious goal that sounds impossibly naïve until you realise he’s already taking concrete steps towards institutional partnerships that could multiply his impact exponentially.

He recently sat in the office of an Executive Officer from the Delhi Development Authority (DDA), discussing a possible internship and the potential to integrate Project Sadak’s data directly into DDA systems.

The goal is a pilot programme where every citizen report on his platform automatically reaches DDA officials for action, essentially making his student-built system part of the government infrastructure.

There are bureaucratic requirements to navigate first. The officer asked Parth to complete an internship and obtain a letter from his school principal, the kind of procedural hurdles that stop most initiatives before they begin.

But for someone who taught himself to build a platform from scratch while juggling schoolwork and research commitments, these feel less like barriers and more like boxes to tick.

The technology roadmap ahead

The current version of Project Sadak is just the beginning.

Parth is planning a WhatsApp bot integration that would allow citizens to report potholes through an app they already use dozens of times a day, eliminating the friction of downloading something new or visiting a website.

“The team is also preparing enhanced AI features that will automatically generate and send reports to authorities without any human intervention, further streamlining the gap between citizen complaints and government action,” he says.

What the platform currently lacks is a monitoring system to track whether repaired potholes remain fixed or deteriorate again.

“There’s no monitoring system as of now,” Parth acknowledges. “But again, people have the platform. If the same pothole appears again, they can report it on the platform.”

It’s imperfect, relying on the same citizens who reported the problem the first time to report it again, but it maintains a cycle of accountability that government systems routinely break.

Confronting the root causes

When asked why India’s pothole epidemic persists despite government apps and entire departments dedicated to infrastructure maintenance, Parth doesn’t soften his assessment.

“I would say that it’s probably due to corruption. Because the quality of the roads, in my opinion, is just not up to the mark.”

It’s a candid diagnosis from a 15-year-old who is actively seeking partnerships with those same government departments.

But rather than letting that realisation paralyse him into inaction, Parth is building the antidote directly into his platform, where every report is public, every email to authorities is documented, and every fixed pothole is photographed with before-and-after images.

Project Sadak isn’t just filling holes in roads; it’s filling holes in accountability and creating the transparency that makes corruption harder to hide and ignore.

Balancing research, school and changing his city

Somehow, while managing Project Sadak’s daily operations and planning its expansion, Parth is also preparing for his presentation in the UK this July, where he’ll discuss his research on AI bias.

His work examines how artificial intelligence systems produce different-quality outputs when asked about Western versus non-Western countries, a technical problem with real-world consequences.

“AI tends to continue on data. If we ask who will be Prime Minister or President of the United States, and then ask about some non-Western country, there’s a higher chance that AI will generate biased information,” he explains.

There’s a fitting irony in someone who is building AI systems to solve local infrastructure problems while also studying the limitations and biases of those same technologies.

Parth plans to pursue computer science with a focus on artificial intelligence at university, but for now, his most impactful work isn’t happening in research papers or academic conferences.

It’s happening on the roads of Delhi, Bengaluru and Gurugram, one filled pothole at a time.

What one teenager proved possible

Project Sadak stands as evidence of a simple truth that’s easy to forget beneath the weight of India’s infrastructure challenges: you don’t need a government position, corporate backing or years of experience to fix civic problems.

Sometimes all it takes is someone who refuses to accept that resignation is the only response available, and someone willing to channel lived experience into systematic change.

In a country where infrastructure problems often feel insurmountable and cynicism feels like the only rational response, a 15-year-old with coding skills and stubborn determination is proving that change starts with someone willing to write the first line of code.

The potholes are still there, far too many of them, but now there’s a platform that actually does something about them.

And that makes all the difference.

Disclaimer : This story is auto aggregated by a computer programme and has not been created or edited by DOWNTHENEWS. Publisher: thebetterindia.com