& Society in Practice

Gabriel Saruhashi's first venture solved a problem that turned out to be only half the problem.

As a Yale undergraduate, Saruhashi majored in computer science and psychology. At SOM, he TA'd for Kyle Jensen's Management of Software Development class – the same class he'd taken the year before. "Getting to turn around and teach it was one of the most fun things I did at Yale," he says. It's also where he picked up the fundamentals he still relies on: how to talk to investors, how to interview real users, how to take a side project from zero to one. "When the cost of writing code approaches zero, knowing what to build and why is the whole job. That's the lesson I carry."

He and his co-founder Zo Orchingwa have now built two ventures on that lesson. The first, Ameelio, gives incarcerated people free ways to stay connected with their families, disrupting an industry that has long profited off prison phone calls. It's scaled to 16 states, reached more than a million people, and drawn funding from Jack Dorsey, Eric Schmidt, and Vinod Khosla, among others.

But Ameelio also showed them the limits of connection alone. "We kept watching people leave prison, having used our tools to stay connected on the inside, and then hit a wall the moment they got out: they could not find a stable career," Saruhashi says. "Solve the communication problem and they're still locked out of the economy."

That wall is what Emerge Career was built to knock down. Emerge trains formerly incarcerated people for high-wage trades – CDL, HVAC, diesel – with an average starting salary above $70,000. New York City Mayor Zohran Mamdani recently announced Emerge's expansion into every sentenced facility on Rikers Island.

Saruhashi frames the opportunity in blunt economic terms: the industries rebuilding America's freight, energy, and grid infrastructure are short workers by the millions, and that shortage is only growing as the current workforce ages out. Meanwhile, formerly incarcerated Americans face unemployment at more than 13 times the national rate. "Those aren't two problems," he says. "They're a mismatch. Emerge is the engine that closes it." He describes the company's larger ambition in similarly sweeping terms: an economic-justice engine for American reindustrialization.

That mismatch shapes how the program actually runs. Students arrive from every point in the system – some still in custody, some on probation or parole, some already home and hitting the same wall again and again. On day one, they're matched to a trade that pays. From there, an adaptive AI engine personalizes their online coursework to their particular strengths and gaps, followed by hands-on training in the shop and the training yard, and finally an AI co-pilot that rides along through the job search itself. The system is built to watch for the moment a student stalls and, crucially, to put a human coach in front of them right then. "So the day-to-day isn't 'talking to a robot,'" Saruhashi says. "It's a real person getting through real training, with software clearing everything in the way that doesn't need human judgment, so the human moments stay human."

That design choice is also a rejection of how workforce development has traditionally scaled. The default strategy in this field is to hire more staff, and Saruhashi argues it backfires on every front: it makes the organization slower to adapt, scales layers of management rather than attention, and scales cost in lockstep with headcount, which is part of why the field spends so much to produce so little. Finally, it leaves quality hostage to which counselor a student happens to draw, with hard-won expertise trapped inside individual heads instead of built into the system.

AI lets Emerge skip both traps: it absorbs the paperwork, reimbursements, and eligibility questions that used to eat up staff time, freeing coaches to show up at the moments that actually decide whether someone finishes. People hear 'AI' and picture a student typing at a chatbot," Saruhashi says. "It's the opposite. The software handles everything that doesn't need a human, which is exactly what frees us up for the parts that do.”

AI also gives Emerge a kind of visibility that used to be impossible. The patterns buried in unstructured data – the texts, the calls, the training performance – once lived only inside a single counselor's head. Now the platform reads across all of it in real time, so the team can see how an entire cohort is progressing as it happens, instead of finding out months later in a report.

When asked about Emerge’s biggest milestone to date, Saruhashi points not to a contract but to the shrinking gap. Federal job centers spend billions to place 38.6% of participants at about $34,708 in first-year pay. Emerge has already delivered hundreds of careers, millions of dollars in public benefit to the communities in which they operate, and outperformed legacy training centers in nine states. This contribution, Saruhashi explains, is their proudest achievement.

And Emerge gets paid only when those outcomes are verified. “Most workforce spending pays for activity regardless of the results,” he says. “We only get paid when someone’s life actually changes.”

Emerge is currently hiring engineers, strategists, and operators. Click here to learn more.