The Career Pipeline Is Broken

Nearly 50% of graduates find themselves underemployed, while 48% confess they don’t feel job-ready.
On the other side, companies wait 35+ days to fill a role, spending thousands per hire as projects pile up.

Worse still, AI automates 45% of entry-level tasks — the very roles students once used to hone their skills are disappearing.

In this broken landscape, the pipe that should carry students into industry has collapsed.
So we asked the question hiding in plain sight: Why force students to beg for jobs, when they could launch as entrepreneurs delivering results?


🚀 Igniting IslaLaunch

IslaLaunch is the rocket built for this broken system — a marketplace where student teams don’t chase jobs, they deliver outcomes.

Operating as DBAs or LLCs, these teams contract directly with companies, offering paid projects as professional services.
Businesses escape the hiring grind, gaining results fast, while students build portfolios, skills, and income.


⚖️ Puerto Rico’s Incentives Fuel the Rocket

  • Young Entrepreneurs Decree: 0% tax on the first \$500,000 for 3 years.
  • Act 60: ≈4% corporate tax for eligible export services.
  • 1099-NEC deductions: Companies expense every dollar under IRS §162.

The result? Teams keep \$0.96 for every dollar earned.
Companies slash onboarding costs. Everyone wins.


🤖 The AI Engine That Powers Flight

  • Lumen → Matches teams to projects with smart embeddings, ensuring perfect fits.
  • Shard → Ranks proposals by expertise, reliability, and past success, making hiring effortless.
  • Prime → Coaches teams with real-time feedback, turning each project into a growth opportunity.

This isn’t just automation — it’s propulsion. Every match, rank, and coaching cycle adds thrust.


🔑 The Impact: Momentum into Orbit

Forged for those who want a different path, IslaLaunch empowers students to become entrepreneurs, not job-seekers.
Companies gain agility, tapping talent without bureaucracy.

And with every successful project, teams gain skills, portfolios, and confidence —
fueling the next launch, compounding momentum, and pushing the whole system into orbit.


🛠️ How We Engineered the Rocket

We built IslaLaunch, a working MVP, in 148 hours with a focus on speed and accuracy.

  • Frontend: Next.js and Vercel handle instant deploys and smooth routing.
  • Backend: Firebase (Firestore + Auth) manages project postings, team profiles, and escrow logic.
  • AI Core: Vertex AI embeddings power project–team matching; a custom pipeline scores proposals on fit, risk, and track record; real-time coaching uses tuned prompts.
  • Forecasting: ARIMA and Prophet test delivery timelines and matching efficiency, feeding insights into scoring.
  • Entity Layer: DBA/LLC contracting integrates Puerto Rico’s Act 60 and Young Entrepreneurs Decree incentives.

Result: A live marketplace where companies post jobs, teams bid, AI selects matches, and contracts are ready to sign — all without manual work.


⚡ Turbulence We Faced

  • Data: No clean dataset, so we simulated workloads to test ARIMA/Prophet.
  • Scoring: Matching worked, but ranking took trial and error to reduce false positives.
  • Entity flows: Translating PR’s tax incentives into contracts revealed tricky edge cases.
  • Pace: Full stack + AI + entity logic in 148 hours forced prioritization over polish.

🏆 Milestones We Reached

  • Built a working MVP in 148 hours, from posting to ranked proposals.
  • Integrated Lumen, Shard, and Prime into a real AI stack — not just prototypes.
  • Used ARIMA + Prophet to feed delivery risk into proposal scoring.
  • Demonstrated how PR’s incentives can support student-run entities.
  • Shipped a marketplace demo where companies post, teams bid, and AI curates.

📚 What the Flight Taught Us

  • Data quality matters — AI is only as good as its inputs.
  • Beyond code matters — contracting and incentives are just as critical.
  • Speed brings focus — the core flow is post → bid → AI curate.
  • Students can be vendors — with the right structure, they’re ready.

🔥 Where the Rocket Goes Next

  • Onboarding: Add initial teams and projects to strengthen AI matching.
  • Payments: Implement escrow with Stripe or a similar provider.
  • Registration: Streamline DBA/LLC setup into a fast, clear flow.
  • Pilot: Launch in Puerto Rico with 10–15 companies.
  • Scaling: Move from prototype models to production-grade AI.

Built With

  • arima
  • cloud
  • firebase-(firestore-+-auth)
  • github
  • google
  • next.js
  • prophet
  • tailwind-css
  • vercel
  • vertex-ai
Share this project:

Updates