Every EV driver knows the feeling: you pull up to a charger, and it’s dead. As I reminded our audience at last week’s Featured Founder webinar, one in five public charging sessions fails today — not because the cars don’t work, but because stations fail silently, and operators don’t find out until a frustrated driver is standing right there. Each fix can mean a technician dispatched at north of $1,000, and two or three trips before it’s solved.

That’s the problem a team of UC Davis and Sac State students set out to solve — and frankly, they’re moving faster than companies ten times their size. Meet Data Pigeon, an AI-native diagnostics platform for EV charging networks. Their models already train on more than 1.4 billion real-world data points and can predict charger failures up to 12 hours before they happen. True to their tagline: “Where EV charging meets reliability.”

Co-founders Senara Millawabandara (CEO), Suhani Shokeen (COO), and Sina Vaghefi (CTO) didn’t start here. The company began as a GovTech idea before a personal frustration redirected them. “Senara actually has an EV, and she’s faced this issue multiple times… every single time she’d go to a charger, it’s always down,” Suhani explained. “The only skill set we could bring was something technical, and that’s what we did. We built machine learning models around charger reliability, and that’s how Data Pigeon was born.”

Rather than chase hype, they went deep. Partnering with the renowned UC Davis EV Research Center, they built charger-specific health baselines and now flag 14 precursor fault codes that historically precede major failures. The metric that matters most to operators? “Our number one metric is mean time to resolution,” Senara said — how fast a broken charger becomes a working one again.

And here’s what I want you to take away: Data Pigeon has customers. They’ve landed their first paying operator — whose highest-revenue chargers in LA now benefit from early warnings — and they’re in active talks with manufacturers eager to integrate predictive maintenance. Their approach is grounded: a free pilot, then a paid partnership once the results speak for themselves. As Sina put it, “Confidence is the key part, and the second part is having a good product.”

When asked what’s next, the answer was simple. “Right now, we want to get more customers,” Senara said — while making the models plug-and-play enough to scale across operators, utilities, and beyond.

If this resonates with you, connect with them. A single warm introduction could be the one that helps Data Pigeon grow. Reach the team at datapigeon.ai or info@datapigeon.ai — and if you know an operator who needs this, reach out to me and I’ll make the intro.

Finally, our gratitude to the partners who make CleanStart’s work possible: the Carlsen Center for Innovation & Entrepreneurship, E2, SMUD, and ChicoStart. Thank you for investing in the founders building our region’s clean future.