For students · Grades 7 to 12

Year-Long Project Mentorship

One student, one project they choose, and six months to a year of personalized 1:1 mentorship to build it, grow it, and show it.

  • Grades7 to 12
  • Length6 months to a year
  • Check-insWeekly, 25 min, hybrid
  • ProjectSolo or a team of up to 4
  • StartRolling
Enroll →How it works
What a year changes

Past the first draft, with someone beside them.

Why one project over six months to a year builds what a short program cannot.

  1. The gap

    Tutoring assigns the task

    The student follows a plan someone else wrote.

    How mentorship closes it

    Their project, their decisions

    The student chooses the problem and answers for the results, with a mentor beside them, not in front.

  2. The gap

    Work only teachers see

    Feedback comes from people who are paid to be kind.

    How mentorship closes it

    Real users, early

    With AI tools, the student builds a working version and shows it to real users early.

  3. The gap

    Short programs end at a first version

    Two days is enough to start, not to make it good.

    How mentorship closes it

    Weekly feedback, for months

    A 25-minute check-in every week, short written updates between, and a term review with families.

  4. The gap

    Applications need real evidence

    A list of activities says little about what a student can do.

    How mentorship closes it

    Five milestones

    Launch, traction, revenue, impact, recognition: the same ladder for every project, judged on the work.

Why it matters now

When AI can make almost anything, what a student chooses, judges, and finishes sets them apart.

Over one sustained project, students learn how a real product is made, how to use the new tools well, and how to keep making their own, for passions, hobbies, and schoolwork. It also gives a college application real, verifiable evidence of initiative. Admissions decisions rest with each university; we do not promise an outcome.

  • Their project, their decisions

    A tutor assigns tasks. Here, the student chooses the problem and answers for the results, with a mentor beside them, not in front.

  • Built for real people, not a demo

    The product goes in front of real users, not a simulated audience. Mentors have built and launched real products themselves.

  • Long enough to get past a first draft

    A two-day workshop ends at a first version. Six months to a year is enough time to test it, revise it, and make it good.

What the research says

Students use AI. Few are taught to use it well.

Three findings from RAND’s national surveys of students. Each links to its source.

How it works

Four phases, one project, at the student’s own pace.

A student mapping problems worth solving
Hong Kong, 2025

Every week, a mentor holds a 25-minute check-in. Between sessions, the student sends a short written update, and families review progress each term.

Milestones

Five milestones, the same ladder for every project.

Each student sets a target with their mentor at the start, then moves up the ladder at their own pace, judged on the work, not on attendance.

  1. 01Launch

    A working product, published at a public link. Every project starts here.

  2. 02Traction

    100 users from outside the student’s own circle.

  3. 03Revenue

    First income from people who pay for the product.

  4. 04Impact

    1,000 users, press coverage, or a competition final.

  5. 05Recognition

    A national competition win, or a registered company.

Where projects continue

Started in a program, kept going.

  • International School of Beijing · After the summer program

    Projects that kept going

    After the five-day program in summer 2026, alumni continue their projects in the Year-Long Project Mentorship.

    Started
    Summer 2026, five days at ISB
    Now
    Year-Long Project Mentorship
    See results →
  • Bodley Academy · Hong Kong, summer 2025

    Twenty students, five days

    A five-day course designed and taught by Phillip An. Each student worked on one real problem they chose.

    Students
    20
    Rating
    9.5 / 10 Net Promoter Score
    See results →
Tools students build with
  • ChatGPT
  • Claude
  • DeepSeek
  • Kimi
  • Figma
  • GitHub
  • Vercel

The student decides what to use, with a mentor checking the result.

What students finish with
  • A live productPublished at its own link, with the student’s name and decisions behind it.
  • A documented recordMonths of users, numbers, and decisions to describe in an application.
  • People who know their workMentors and guest reviewers who watched them build it.
  • Experience presenting itPractice explaining the work, in public, to people who ask hard questions.
Our team

Our team has walked the same path.

Arena’s team studied at and worked for the institutions students aim for, and teaches from that experience.

Studied at
  • Caltech
  • Harvard University
  • Brown University
  • Tsinghua University
  • International School of Beijing
Worked at
  • McKinsey & Company
  • Goldman Sachs
  • Y Combinator
  • Apple
  • Microsoft AI
  • Deloitte
  • Kimi (Moonshot AI)
Arena leadership
Meet our leadership →
Questions

What families ask.

What if my child doesn’t have an idea yet?

That is where most students start. In the first phase, the student and mentor explore problems the student cares about and test which one is worth building.

Does this fit a busy schedule?

Yes. The weekly check-in is 25 minutes, and a written update between sessions takes a few minutes. The program is designed to sit alongside exams, sports, and other commitments.

Does my child need a team, or can they work alone?

Either works. Teams of up to four are encouraged, and students may also work on their own. Students can regroup at the end of a term.

Will this help with college admissions?

No program can promise an admissions outcome, and we do not. What a student gains is a real project with users, results, and decisions they can describe in their own words and show with a link. Each university makes its own decision.

What does it cost?

It depends on the project and the length of mentorship. Tell us about your child and their project idea, and we will confirm details directly within three working days.

Can a student join without a project idea?

Yes. The first phase is choosing and validating the problem with their mentor.

Next step

Enroll at any time.

Mentorship starts on a rolling basis. The form takes two minutes, or ask us a question first.

Enroll →Ask a question first
A mentor reviews a student’s product on a laptop
Reviewing a student’s product, Beijing, 2026
Two ways to join
  1. 01
    Continuing from ArenaArena alumni

    Mentors continue from a student’s earlier Arena project.

  2. 02
    Starting freshNew to Arena

    The student chooses a project with a mentor, then check-ins begin.

Preview of the Arena School one-pager

Get the Arena one-pager

Every program and what we have run, on one page.

See all Arena programs →