Paying Teens to Work Before College
Deep24 has launched a gap‑year program that pays recent high‑school graduates six‑figure stipends to work inside its portfolio of AI‑driven autonomous businesses. Oliver Zou, who left the University of Pennsylvania a year early and worked at Uber, founded Deep24 in 2024 and now runs the initiative through his nonprofit Common Gap. The program matches graduates with full‑time roles at early‑stage AI startups; each host must pay at least a year, and Zou says most offers exceed . Common Gap takes no equity, charges no fee, and simply passes the salary from startup to teen.
Zou says he wishes he had taken a gap year himself. “Doing real work, not just turning in assignments for a grade, feels rewarding and teaches faster,” he explained. He sees the gap year as a deliberate choice, not a default path to college.
How the Program Matches Teens with Startups
Applicants answer a single prompt: “the most impressive thing you’ve built or done.” An optional video interview follows. Zou is testing a language‑model screen for early rounds, but he calls it a speed‑up, not a gate. The name echoes the Common App, the college‑application platform used by over a thousand schools.
Deep24, the productivity analytics startup he created, ranks among the first hosts. Other launch partners are SuperBuilders in Austin (which builds tech for Alpha School, an AI‑powered private school), a health‑tech startup, and an insurance‑focused company. All agree to the minimum stipend; Zou anticipates most offers go higher.
Placement works like a mentorship apprenticeship. Each teen receives a dedicated engineer or product lead from the host company. Weekly check‑ins, code reviews, and invites to product‑strategy talks replace generic onboarding. The teen’s project shapes the learning path; if they troubleshoot data pipelines for Deep24’s time‑management tool, the mentor points them to relevant Stack Overflow threads, docs, or peers.
Selection favors “high agency” over grades. Interviewers ask candidates to describe a project they owned, how they tackled unknowns, who they consulted, and what they learned solo. Zou’s own path — leaving Penn early, working at Uber, and building Deep24, shows he values learning by shipping real products with real consequences.
Why Startups Pay Six Figures for Teen Talent
Offers for top machine‑learning roles have risen this past summer. At the same time, tech layoffs have topped nearly 50,000 since 2022, yet frontier labs and well‑funded application‑layer firms continue to hunt early‑career engineers aggressively.
Zou frames the gap year as a supply‑side response: “Why wait for them to graduate?” He argues a 19‑year‑old who has already shipped code can contribute to a product roadmap on day one, whereas a junior hire from a campus pipeline often needs months of ramp‑up.
For early‑stage AI companies, paying the minimum stipend feels like a feature, not a bug. Two gap‑year engineers paid at the top of the scale each deliver about 4,000 engineering hours per year — roughly the output of two full‑time engineers. That time can move a prototype from notebook to API, run the A/B tests that unlock a seed round, or close compliance gaps that block a health‑tech launch. The alternative — leaving the role empty or hiring a single senior engineer, either stalls the roadmap or blows the burn budget.
| Description | Amount |
|---|---|
| Minimum stipend (host requirement) | $75,000/year |
| Typical offers (most exceed) | >$100,000/year |
| Top machine‑learning offers | high‑six‑figure range |
| Senior engineer alternative | $300,000+/year |
| Thiel Fellowship | $200,000 (lump sum) |
Critics worry the model compresses adolescence into a high‑stakes labor market. Zou counters that participants who later choose college do so “very intentionally,” with cash savings and a portfolio of shipped work. For startups, the calculation is simple: they get technical leverage now, at a price their runway can absorb, from a talent pool the rest of the industry still waits to graduate.
What Universities Say About the Shift
University career centers, built on multi‑year degree pathways, see the gap‑year‑to‑six‑figures model as a recruitment shortcut that squeezes the timeline between education and employment. Traditional internship providers warn it threatens the scaffolding they have assembled over decades.
Oliver Zou states explicitly that he bypasses conventional gatekeepers. Business Insider reports that the application skips résumés entirely and asks for “that prompt,” with an optional video clip. He is also weighing AI as an initial filter for early interview rounds. That design — favoring demonstrable output over pedigree, mirrors a broader shift already visible in Silicon Valley.
Venture‑capital precedents give the model a veneer of legitimacy. Kleiner Perkins runs a Common Gap program that places college students in portfolio‑company internships, though it does not guarantee six‑figure pay. The Thiel Fellowship, which began in 2011, awards to founders who drop out of college.
Some campuses are adapting. Vanderbilt announced a New York City campus in September 2024 that opened applications for fall 2026. The City University of New York’s CUNY Beyond initiative aims to triple paid internships by 2030. Manhattan College reports 87% of alumni work in their chosen field, often at firms like IBM and JPMorgan. These programs treat internships as stepping stones, not endpoints, and anchor that philosophy in degree completion.
Faculty Development Day at the New York Institute of Technology will examine how generative AI reshapes teaching and assessment — a reminder that traditional academia still sees itself as a counterweight to purely commercial training.
The Wider AI Talent Race
The contest for early AI talent is widening beyond pure salary wars. A September 2025 counter‑moves roundup shows firms dangling NBA‑style pay packages that reach tens of millions of dollars to lure scarce expertise. Research on Klover.ai from July 2025 details how Linde, the industrial‑gas giant, boosts its talent strategy. Linde invests in upskilling its global workforce and rents elite specialists from partners such as appliedAI and DAIN Studios, letting top AI talent tackle real‑world physical problems. AI and digital initiatives drove about one‑third of Linde’s productivity gains last quarter, lifting plant throughput by 4% and cutting energy use by 2%. Those gains translate into savings of several million euros each year, with a payback period of six to twenty‑four months.
These outcomes explain why securing talent early compounds advantage. The recent surge in top‑machine‑learning offers — already noted for reaching very high levels, pushes companies to look before the traditional college pipeline.
Deep24’s approach, which pays teens at the top of the stipend range, shifts the focus from college dropouts or enrolled students to high‑school graduates, broadening the pool of early‑career entrants into AI‑driven ventures.
Whether the trend accelerates or settles will hinge on how quickly other studio models copy the structure and whether academia can reshape its early‑career offers to match the stipend levels and project exposure on offer.
Expanding to San Francisco and Beyond
Oliver Zou says he is shaping the next phase of Common Gap by bringing additional autonomous businesses into the program. He is arranging placements with startups based in San Francisco, aiming to deepen the studio’s presence in a major AI hub. This step builds on the current group of partners, all of which commit to paying participants at least the minimum stipend.
Zou is also testing an artificial‑intelligence screen for early applications. The tool would review the “that project” submissions and optional video clips before human reviewers see them, cutting the manual work needed to evaluate a growing pool of candidates. By automating this first step, the program can handle more applicants and more host companies without adding staff.
Together, the geographic push and the AI screen form a growth strategy for the gap‑year effort. More host companies mean more paid positions for teens, which in turn draws more applicants. This cycle could help Deep24 staff its ventures faster and stay competitive for the same high‑school talent that traditional internships and university career offices pursue. The coming months will show whether the San Francisco expansion yields concrete deals and whether the AI filter becomes a regular part of the selection process.
A Day in the Life of a Gap‑Year Engineer
The premise is simple: host companies agree to pay at least the minimum stipend, and Zou anticipates most offers will go higher. This shifts a gap year from a break with no income to a source of earnings that rivals many early‑career salaries.
In the Common Gap network, daily work centers on real product contributions rather than routine tasks. One participant at a San Francisco‑based AI startup described a day that starts with a stand‑up over the team’s async channel, then moves to benchmarking prompt performance across the company’s language models. The teen logs results in a shared spreadsheet that engineers review before weekly releases. Though the task fits within a sprint, its substance earned a mention in the product‑roadmap meeting from the startup’s CTO — exactly the kind of integration Zou had in mind when he called the experience “that idea.”
A different track places teens with founders building AI‑driven autonomous businesses. Here the work often begins with gathering and labeling sensor data from delivery fleets. Mornings are spent organizing logs; afternoons involve writing Python scripts to clean and categorize the information; evenings are devoted to shadowing a senior engineer’s code review.
The stipend — set at the program’s floor, with most participants negotiating toward the higher end, covers full‑time hours, and the teen said the pay let them treat the year as a paid apprenticeship rather than a pause.
Placement works like a matchmaking service. A single form funnels applicants to multiple startup interviews, cutting the repeat‑apply hassle. Zou launched the idea after seeing the gap between elite machine‑learning offers that reach very high levels and high‑school graduates who can build real tools. The “that” question sparked the program. Outcomes show teens splitting time between two hosts or staying a full year with one, depending on the host’s needs and the learner’s goals.
Critics worry that large sums for 18‑year‑olds speed up pressure on traditional internship tracks. University career centers watch enrollment shifts, concerned that a high‑school earner making top‑tier stipend pay could shrink the appeal of short, low‑paid college internships. Zou acknowledges the tension, noting that college isn’t the right path for everyone and has become the default choice for many. He says the program offers a deliberate alternative for those who want to test whether classroom learning matches their real interests, while letting them earn a wage that many graduates only see after years of entry‑level work.
What links these stories is the lack of busywork. Whether the teen is benchmarking language models, sorting sensor logs, or writing data‑cleaning code, the output leaves the laptop and enters the startup’s workflow. The guaranteed floor, with expectations of exceeding it, ensures the teens receive a wage comparable to many post‑grad offers.
Because the startups fund the salaries directly and the initiative runs as a non‑profit, incentives align: companies obtain trial talent, teens receive paychecks, and the route from high‑school exit to AI‑studio employment shrinks from years to months.
The early version, validated by Y Combinator’s Silicon Valley program, shows the model passes a tough test. Zou’s plan to add an AI screen for early interviews points to a more automated selection as the program grows. For current participants, the experience delivers both a paycheck and proof of skill in a market that once reserved such chances for college graduates with internship backgrounds. Whether they return to school or double down on autonomous ventures, the gap year functions as a genuine on‑ramp, one that Deep24’s broader studio effort and initiatives like Common Gap are pushing to legitimize.
A teen’s laptop screen flickers with a newly shipped model, the stipend check tucked beside it, signaling that the gap year has become a paycheck‑powered prototype.
Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.