Document work a student can genuinely explain and defend

A portfolio is not an admissions object with a required project count. It is an evidence system for work the student owns and can explain: dates, role, question, new contribution, methods, student-authored code, data provenance, results, limitations, iterations, verification, artifacts, and mentor disclosure. Students may begin without programming experience, but they must understand and explain the tools and code used in the final work.

No university-wide project count, duration, or competition record is implied. Last reviewed: July 11, 2026.

Why an AI portfolio matters in admissions

A portfolio records what a student built, the role they played, the evidence behind the result, and what they learned. Whether it may be submitted in an application depends on the university and route, but a well-documented process also supports learning, interviews, collaboration, and project handoff.

A good portfolio isn't a list of finished results. It has to tell a story—the motivation behind finding the problem, the trial and error of solving it, the reasons behind technical choices, and how you grew through the project. At CIT, we build out the way to systematically design and document this whole process together with you.

CIT Evidence Packet — Reviewed July 2026

For each substantial project, record the title and dates, exact student role, question, coursework anchor, new contribution, methods and tools, student-authored code or work, data provenance, results, limitations, iterations, verification, report/demo/repository, and mentor support. One project may create several artifacts; those artifacts are not automatically separate activities.

Source: aimi.stanford.edu · inspiritai.com · veritasai.com

What belongs in a verifiable AI portfolio?

No university publishes a universal project count, duration, or competition formula. The useful standard is whether the student can explain their exact role and verify the work truthfully.

Evidence item Document Verify Reporting rule
Question and new contribution Summary and log Sources and requirements Separate from assessed work
Student-authored code, demo, or report Artifact Version and authorship Check privacy and permissions
Student role and mentor support Role record School, team, mentor Do not overstate
Data, methods, results, and limitations Reproducibility Baselines and tests Include failures
Official result or verification Link evidence Official record Check route before use
Projects and artifacts No fixed count Ownership and continuity Do not split one activity artificially

Can students who can't code still build an AI portfolio?

An AI portfolio isn't just for students aiming at a CS major. Students aiming for medicine/life sciences, humanities/social sciences, business, or the arts can also stand out in admissions with an original project that applies AI tools to their own area of interest. The table below lays out project directions and key AI technologies by major.

Target major/field AI-fusion project direction Key tools & technologies
Life sciences/medicine (Pre-med) Medical image classification, drug-molecule property prediction, public-health data visualization Google Teachable Machine, Python (pandas), Kaggle datasets
Humanities/social sciences Sentiment analysis of historical texts, social-media opinion analysis, AI ethics policy research reports Hugging Face sentiment classification, ChatGPT API, public data portals
Business/economics Automated consumer-review classification, financial-data prediction models, startup pitch-deck automation Scikit-learn, Tableau, Google Colab (no-code flow)
Arts/design Generative-AI art projects, music-emotion classification, digital-media accessibility tools Stable Diffusion, Magenta (Google), p5.js + ML5.js
Environment/earth sciences Satellite-image-based deforestation detection, climate-data visualization, carbon-emissions prediction Google Earth Engine, NASA open data, Python visualization

CIT matches mentors across the fields above. Students may begin without programming experience, then learn the coding required for the chosen project. They must understand and explain every tool, AI output, line of code, and analysis used in the final work. See the full AI program curriculum →

1:1 online mentor matching for students pursuing non-CS majors

Using the same curriculum as our in-person classes in Apgujeong, we run 1:1 online sessions with dedicated mentors for the life sciences, humanities, business, and arts. Korean students living in Singapore, Hong Kong, or the U.S. East Coast, as well as families who find it hard to travel within Seoul, can receive the same level of portfolio guidance.

How online sessions are structured
  1. 1. First session: mapping interests, intended major, prior learning, and assessed-work boundaries; the student chooses the question
  2. 2. Sessions 2–4: hands-on data collection and applying AI tools
  3. 3. Sessions 5–8: refining the deliverable + writing the GitHub README
  4. 4. Wrap-up: organizing a truthful evidence packet and route-compatible presentation materials
This is a good fit if you are
  • A student who may be new to coding and is ready to learn the level required for the chosen project
  • A Korean international-school student living abroad (Singapore, Hong Kong, the U.S.)
  • A student targeting pre-med, life sciences, humanities, or arts who wants a major-aligned, defensible project
  • A Grade 9–11 student with enough time to build authentic work without compromising school priorities

CIT's portfolio design process

Step 1: Choosing a project topic

CIT helps map the student's interests, intended major, prior learning, available time, and assessed-work boundary. The student chooses the independent question and remains responsible for every project decision.

Step 2: Building and executing

Students carry out a real project—implementing an AI model, building an app, analyzing data, and more. Alongside technical mentoring, the development process is documented systematically, and the trial-and-error and improvement process is treated as an important part of the portfolio.

Step 3: Documenting and organizing

Students write their own project report, GitHub README, and presentation materials. CIT teaches structure and gives permitted feedback on clarity, reproducibility, and whether the student's exact role, methods, results, limitations, and support are disclosed.

Step 4: Evidence packet and route-specific use

We organize the project log, code, data provenance, demo, report, and verification. If an admissions route permits the material, students describe their exact role, dates, outcome, and mentor support truthfully.

Where the portfolio is used

U.S. and U.K. applications: A route may allow a student to describe the project in activities, additional information, or an interview. The description must state the student's exact role, dates, outcome, limitations, and mentor support. GitHub links, apps, and competition results are not universal requirements. See the full EC strategy →

Korean university routes: Whether a portfolio, outside competition, award, or repository may be submitted—and whether school verification is required—depends on the university, admission year, and route. The current official guide takes priority.

School activity: A project may become a school-approved club, service, or research presentation when appropriate. Records must reflect the student's real role and verified outcome.

Your portfolio starts with a single 'AI project'

There is no required project count. Start with one substantial project the student owns and can explain, then document the question, role, data, methods, results, limitations, and verification. The study guides below are learning resources; assessed coursework is not copied into a new project.

Frequently Asked Questions

What should go into an AI portfolio?

A portfolio records a problem the student found, the solution process, a working result, the exact student role, data provenance, iterations, limitations, verification, and mentor support. There is no required project count, and multiple artifacts from one project are not automatically separate activities.

Can students who can't code still build an AI portfolio?

Students may begin without programming experience. The coding required depends on the project, and students must understand and explain every tool, AI output, line of code, and analysis used in the final work. CIT offers 1:1 mentoring across life sciences, humanities, business, and the arts.

What belongs in a verifiable AI portfolio?

Record the project title and dates, the student's exact role, question, coursework anchor, new contribution, methods and tools, student-authored code or work, data provenance, results, limitations, iterations, verification, artifacts, and mentor support. No university publishes a universal portfolio formula.

How many AI projects do I need?

There is no required project count or fixed duration. One sustained project may produce a report, poster, demo, repository, presentation, and award, but those artifacts are not automatically separate activities. Prioritize student ownership, relevance, truthful verification, and the ability to explain every decision.

What grade should I start building a portfolio?

Timing depends on the student's prerequisites, current coursework, examination and assignment deadlines, and intended output. Younger students can build foundations and explore; older students should extend one strong existing area narrowly. Academic performance and required schoolwork come first.

Can I build a portfolio without any competition wins?

Of course. Competition results are just one element of a portfolio. A project where you found and solved a problem on your own, work you kept developing over time, and your role in a team project can carry even greater value.

Does a coding portfolio help with U.S. college admissions too?

Student-owned coding and AI work may provide evidence of interests, technical skill, and initiative. GitHub, apps, and competition records are not universal requirements. If a route permits the material, describe the student's exact role, dates, outcome, limitations, and mentor support truthfully.

Do you help build a GitHub portfolio too?

Yes. We coach students on setting up a GitHub profile, writing READMEs, documenting projects, and managing their commit history. Especially when applying to U.S. universities, a GitHub link becomes an important piece of evidence of technical ability.

How long does it take to design a portfolio?

Timing depends on the state of the existing work, code and data permissions, new validation required, and the academic calendar. We plan in relative weeks after diagnosis and reduce or pause work around mocks, examinations, and required coursework.

Can I join the CIT portfolio program even if I don't live in Apgujeong?

Yes. CIT runs 1:1 online sessions using the same curriculum as our in-person classes in Apgujeong. For students pursuing non-CS majors in particular, we match you remotely with a dedicated mentor in the life sciences, humanities, business, or arts. Korean students living in Singapore, Hong Kong, and the U.S. East Coast are already taking part. For details, see the online class guide page.

Consultation info

Not sure how to get started on a portfolio? In a free consultation, we'll walk you through a portfolio strategy tailored to the student's current level and goals.

Related Pages

References (Sources)

  1. Stanford AIMI — aimi.stanford.edu
  2. Inspirit AI — inspiritai.com
  3. Veritas AI — veritasai.com
  4. Common App Activities Section — commonapp.org

ACADEMIC → INDEPENDENT EXTENSION → VERIFIABLE EC

Connect academics to a defensible EC

Protect grades and required school submissions first. Then branch into a new, student-owned question and document evidence that the applicable route permits.

Course knowledge + new question + new evidence + student ownership + verification = defensible EC

  1. LearnDevelop subject knowledge and coding, statistics, and research methods through IB, AP, IGCSE, A-Level, or school courses.
  2. Secure the academic outcomeProtect grades, exams, predicted results, and required school submissions first.
  3. Set the integrity boundaryRecord the boundary between submitted or assessed code, data, and writing and the new work.
  4. Branch into a new questionAdd a substantive new question, dataset, method, user group, experiment, or outcome.
  5. Build and validateThe student creates the code, experiment, analysis, and research log and explains the limitations.
  6. Externalize the workWhere appropriate, connect the project to KSEF, a school activity, competition, CAS or service, real-world deployment, or a portfolio.
  7. Document and adaptCreate a truthful evidence packet and adapt it only for admissions routes that permit it.
Extend the learning; do not duplicate the assessed submission

CIT helps students extend what they have learned into new work. We do not duplicate assessed submissions, write school coursework for students, or present old work as a new competition project. The student must make the decisions, create the work, and be able to explain every part.

  • Green — normally reusable: These may form the foundation of an independent extension.
  • Yellow — review required: Use only when permitted, disclosed, and clearly distinguished through a substantive new contribution.
  • Red — do not reuse as a new submission: These must not be submitted as a separate original work.
View the full Coursework-to-EC pathway → Check route-specific Korean admissions evidence →

CIT does not duplicate assessed submissions, write student coursework, or guarantee awards, international selection, or admission.

Get a Consultation (02) 540-2922