Just 18 months ago, Karim Jouini and Jihed Othmani had crossed the finish line of the startup marathon. Their fintech company, Expensya—an expense management platform born in Tunisia—had been acquired by Swedish software giant Medius in a deal reportedly worth over $120 million. With the sale widely regarded as one of Africa’s largest tech exits, the duo seemingly had every reason to step away from startup life.
But startup life, as it turns out, wasn’t done with them.
Today, Jouini and Othmani are back in the founder’s seat, this time leading a new venture: Thunder Code, a generative AI-powered software testing startup. Headquartered in Paris with an office in Tunis, Thunder Code has already raised $9 million in seed funding, drawing interest from top-tier backers like Silicon Badia, Janngo Capital, Titan Seed Fund, and strategic angels including Station F’s Roxanne Varza and InstaDeep’s Karim Beguir.
The idea for Thunder Code emerged not from boredom, but from Jouini’s firsthand experience post-acquisition. As Chief Technology Officer at Medius, he oversaw the integration of six acquired companies across three continents. It was during this role that he began to see a universal and persistent pain point: the inefficiency of traditional software testing.
“Testing was always the bottleneck,” he recalled. “No matter how innovative a product was, teams struggled with manual QA processes. It was slow, repetitive, and expensive.”
Thunder Code aims to change that narrative with autonomous AI “agents” that simulate human testers. These agents are capable of navigating complex user interfaces, generating test cases, catching bugs, and learning from past behavior to continuously improve. The promise? A 90% reduction in testing time, freeing developers to ship faster with greater confidence.
What sets Thunder Code apart isn’t just its use of generative AI—it’s the founders’ renewed obsession with speed and precision. “We built and shipped our MVP in just six weeks,” said Jouini. “By six months, the product was already more robust than Expensya was in year four.”
That speed isn’t accidental. Having learned the hard way during their first startup, the co-founders are building Thunder Code differently. From hiring top-tier talent early to being open to equity dilution in exchange for growth, they’re applying every lesson from their past. “A lot of African founders fear dilution,” Jouini explained. “But if giving up equity helps build a billion-dollar business faster, that’s value.”
The startup is already generating traction with pilot programs and paying clients in the U.S., France, Tunisia, and Canada. While currently focused on web application testing, Thunder Code plans to expand into mobile, desktop, and API testing by late 2025—chasing a global market projected to surpass $100 billion in value.
While the software testing landscape is crowded—with incumbents like Tricentis and BrowserStack, and new players like Nova AI and Jetify—Jouini believes Thunder Code’s execution speed and AI depth give it an edge. That confidence is bolstered by the technical prowess of co-founder Othmani, who had already been developing internal AI tools at Expensya years before ChatGPT went mainstream.
Perhaps most inspiring is the startup’s ecosystem. Some of the early backers are former Expensya employees who cashed out during the acquisition and are now reinvesting in the founders’ second act. “They believed in us then, and now they’re part of the next chapter,” Jouini noted with pride.
Thunder Code isn’t just a rebound project. It’s a vision shaped by experience, sharpened by setbacks, and powered by a belief that the future of software development hinges on faster, smarter, AI-driven tools. And for two founders who once promised never to start another company, they’re moving at lightning speed.
