OCI-hackathon: OpenPath

A career-discovery tool for students who aren't sure what they want to do.

OpenPath helps students uncover a wide range of career options, including specialized and less familiar job titles, before they spend money or time on a major, classes or training. It is a practical exploration tool: it suggests paths to look into and points to current sources to verify them. It does not predict where anyone will thrive, and it is not a complete catalog of every job.

Live demo: https://hackathonvanguard.vercel.app

Example result in one click: https://hackathonvanguard.vercel.app/?demo=maya (a sample student's finished match and plan; add ?demo=maya to any copy of the app)

Repository: https://github.com/ble-fortinet/OCI-Hackathon

Who it's for

• High school and college students choosing classes, majors or training
• Career changers exploring adjacent fields
• Counselors, teachers and parents who want a starting point for career conversations

How to run it

OpenPath is one self-contained file. No account, server or install is needed.

Optional AI explanation. When deployed on Vercel with AI credentials, the results page shows "Explain this match with AI", served by api/advisor.js. It uses OpenAI when OPENAI_API_KEY is set, otherwise OCI Generative AI when all OCI values are set. Add the variables in the Vercel project (Settings → Environment Variables), then redeploy. The live demo currently uses OpenAI. When a student presses the button, only the career title and the scenario labels they answered yes or maybe to are sent. The server looks up the career's facts in api/_careers.json (generated by tools/build.py), checks every number in the AI's answer against those facts, and limits each visitor to 10 explanations per 10 minutes (AI_LIMIT_PER_IP) and each server instance to 300 per hour (AI_LIMIT_GLOBAL). A generated explanation is included in the printed plan. No name, account or free-text note is sent.

OpenAI: set OPENAI_API_KEY, and OPENAI_MODEL if your project can't use the default gpt-4o-mini (for example gpt-6.1-sol).

To test a key locally: copy .env.example to .env.local, fill it in, run npm install, then node --use-system-ca tools/test-oci.js.

OCI Generative AI variables:

Variable	Value
OCI_TENANCY_OCID	Your tenancy OCID
OCI_USER_OCID	The OCID of a user with access to Generative AI
OCI_FINGERPRINT	The fingerprint of that user's API signing key
OCI_PRIVATE_KEY	The full PEM private key
OCI_COMPARTMENT_OCID	A compartment where the user may use Generative AI
OCI_REGION	Optional. A region that offers OCI Generative AI. Default us-chicago-1
OCI_GENAI_MODEL	Optional. Default meta.llama-3.3-70b-instruct

The user needs an IAM policy such as allow group <group> to use generative-ai-family in compartment <compartment>. Without these variables the button stays hidden and the app works as before.

• Open it directly: double-click index.html, or open it in any modern browser.
• Deployment: every push to main on https://github.com/ble-fortinet/OCI-Hackathon deploys automatically to https://hackathonvanguard.vercel.app via Vercel.

Quick match (the default)

Interests and work styles are shown as icon tiles (version 3); each question card shows the icon of the area it comes from.

For students who want a fast answer, Quick match narrows to one top match in a few taps:

1. Pick the interests that sound like you from the 10 icon tiles (as many as you like; fewer gives a sharper match), and optionally any work styles.
2. Answer short "Would you enjoy this?" questions with Yes, Maybe or Not for me, one at a time. Each question is a realistic scenario chosen to best split the careers still in the running, so most students finish in 6–12 taps. Back, Skip, "Change my picks" and "Show my match now" are always available.
3. See your top match first, with a full step-by-step path, then 5 other strong matches. Tap any of them to open its own path.

Scoring. Each career gets +3 for every scenario answered "Yes" that it fits, +1 for "Maybe" and −3 for "Not for me". It also gets +1 for each chosen interest and +0.5 for each chosen work style it connects to. Careers with role-specific research get a tiny tie-breaker. Until the student says yes to something, the first questions rotate through the picked areas (all 10 if none are picked), trying a different kind of work each round, so every area is probed early. After a "yes", the next questions favor the same kind of work and area. A top match is shown only after at least two "yes" or "maybe" answers (or after 20 questions); then questions stop when there's a clear leader (at least 5 answers and a 3-point lead, or a single contender), after 8 answers with any lead, or at 12 answers. Students can always answer 4 more to sharpen the result.

The step-by-step path appears for every match and at the top of every career page:

1. Now, in high school: classes, student organizations and low-cost ways to try it.
2. What to study: the most common education level, majors and programs, useful courses, and what to check before paying for a program.
3. First 0–2 years: the first role, credentials to look into, and starting pay.
4. Years 2–7: skills to build, networking, mid-career credentials and typical pay.
5. Years 8+: leadership, specialization or pivot options, advanced credentials and senior pay.

Print or save the plan. "Print my plan" (on the top match, each other match and every career page) opens a clean plan view with Print and Download plan buttons. The download is a standalone page that opens in any browser and prints well, which helps where printing is blocked (some embedded previews block it).

The top match is the best fit for the student's answers, a starting point to explore, not a guarantee or a prediction.

Detailed quiz (hidden for now)

The detailed quiz from version 1 is hidden in this version. Its code is still included: to show it again, open tools/template.html, change const SHOW_DETAILED_QUIZ = false; to true, and run python tools/build.py. When shown, it works like this:

The quiz is the main way in. Students can pick more than one answer at every step and go back to change anything without restarting. Answers are saved in the browser.

1. Interests – 10 broad areas: people and culture, analysis and problem-solving, creative work, community, operations, science, communication, leadership, environment and technology.
2. Kinds of work – for each selected interest, 4–7 more specific kinds of work (for example, under analysis: finding patterns in data, troubleshooting machines, investigating why something happened, solving a customer's problem, working through a scientific question, puzzles with money and risk).
3. Scenarios – for each selected kind of work, realistic tasks from real jobs (for example "Diagnosing why a robot or machine keeps stopping" or "Spotting an unusual pattern in a dataset that nobody noticed").
4. Work style (optional) – hands-on work, teamwork, clear structure, fast-changing work, outdoors or fieldwork and independent focus. Each has its own specific choices and scenarios.
5. Your note (optional) – a free-text box. It only does basic word matching against career titles and descriptions, so guided answers give clearer matches. The note stays in the browser.

Matching rules. Each career card is tagged with the scenarios it fits. A career appears when it connects to at least one of the student's most specific choices: a selected scenario, a selected kind of work with no scenarios picked under it, or a selected interest with no kinds picked under it. Picking several things widens the results so students can explore combinations rather than being locked into one family.

Ranking. Each connection adds points: interest 1, kind of work 3, scenario 6; work style 1, specific 2, scenario 4; each matched note word 2 (up to 3 words). Every result card lists the scenarios, kinds of work, interests, work styles and note words that connected it. Results can be filtered by career family ("All paths" or one family) and sorted by strongest connection or A to Z.

Career coverage

319 unique career-title cards across 19 families:

• People, HR and workplace culture: 20
• Community and public service: 14
• Visual art and design: 16
• Film, TV and animation: 16
• Games and digital content: 11
• Music and audio: 9
• Writing and communication: 15
• Technology, data, AI and cybersecurity: 26
• Robotics, automation and manufacturing: 11
• Engineering and skilled trades: 20
• Healthcare, physicians and medical specialties: 39
• Science, research and space: 18
• Education and learning: 16
• Law and public safety: 13
• Hospitality, events and tourism: 9
• Agriculture and animal care: 12
• Transportation and aviation: 12
• Sales, business and operations: 25
• Environment, climate, conservation and energy: 17

This is a starting set, not every job. Titles vary by employer, region and time. Each card links to an official occupation where one exists, and the About page points to O*NET OnLine for anything not here.

Role-specific profiles vs shared family guidance

97 careers have role-specific research: a typical day, work environment, skills, credentials to verify (with when they usually fit), courses, high school classes, majors, networking actions, ways to get experience as a student, a five-stage path and specialize/lead/pivot moves, plus a researched pay range with a source link. These pages are labeled "Role-specific details."

• People, HR and workplace culture: Benefits Specialist, Career Coach, Compensation Analyst, Corporate Trainer, Employee Engagement Specialist, HRIS Analyst, Inclusion Program Specialist, Onboarding Specialist, Organizational Development Consultant, People Analytics Analyst, Talent Acquisition Coordinator, Workforce Planning Analyst
• Community and public service: Development Officer, Peer Support Specialist, Policy Advocate, Victim Advocate
• Visual art and design: Exhibit Designer, Food Stylist, Medical Illustrator, Type Designer
• Film, TV and animation: Colorist, Foley Artist, Motion Designer, Set Dresser
• Games and digital content: Community Manager, Narrative Designer
• Music and audio: Luthier, Music Therapist, Podcast Producer
• Writing and communication: Captioner, Conversation Designer, Copywriter, Court Interpreter, Grant Writer, Localization Specialist, Public Affairs Specialist, Technical Writer, UX Writer
• Technology, data, AI and cybersecurity: Accessibility Specialist, Cybersecurity Analyst, Data Visualization Specialist, Digital Forensics Examiner, GIS Analyst, QA Tester, Scrum Master, Solutions Engineer, Trust & Safety Specialist, UX Researcher
• Robotics, automation and manufacturing: Drone Pilot
• Engineering and skilled trades: Biomedical Equipment Technician, Elevator Mechanic, Model Maker, Theatrical Rigger
• Healthcare, physicians and medical specialties: Audiologist, Child Life Specialist, Clinical Research Coordinator, Genetic Counselor, Health Educator, Medical Scribe, Patient Navigator, Perfusionist, Prosthetist / Orthotist, Sleep Technologist, Speech-Language Pathology Assistant
• Science, research and space: Astronaut, Food Scientist, Hydrologist
• Education and learning: Academic Advisor, Archivist, Instructional Designer, Museum Educator, Museum Registrar
• Law and public safety: Compliance Analyst, Fraud Analyst, Mediator, Paralegal
• Hospitality, events and tourism: Event Producer, Tour Manager
• Agriculture and animal care: Animal Enrichment Specialist, Wildlife Rehabilitator
• Transportation and aviation: Avionics Technician
• Sales, business and operations: Actuary, Brand Strategist, Customer Success Manager, E-commerce Specialist, Medical Sales Representative, Procurement Specialist, Real Estate Appraiser, Revenue Operations Analyst, Sponsorship Coordinator, Sports Agent, Supply Chain Analyst

The other 222 careers use shared family guidance plus official occupation data. Their pages are labeled "Shared family guidance + official occupation data" and say clearly that they are not detailed research about that exact title. They combine:

• general guidance written for the career family (environment, low-cost ways to explore, networking, ways to practice, long-term growth), and

• data for the linked official occupation: ONET tasks, related titles, skills, tools, education levels and suggested classes; NCES-linked majors; BLS pay; and possible next moves computed from ONET skill and task similarity.

• People, HR and workplace culture (20 careers): Executive Coach, HR Business Partner, HR Generalist, Human Resources Manager, Labor Relations Specialist, Payroll Specialist, Recruiter, Training and Development Manager

• Community and public service (14 careers): Child, Family and School Social Worker, Community Health Worker, Eligibility Interviewer, Emergency Management Director, Nonprofit Program Manager, Probation Officer, Recreation Coordinator, Rehabilitation Counselor, Substance Abuse Counselor, Urban Planner

• Visual art and design (16 careers): Architect, Art Director, Fashion Designer, Floral Designer, Graphic Designer, Industrial Designer, Interior Designer, Landscape Architect, Museum Curator, Photographer, UX/UI Designer, Visual Merchandiser

• Film, TV and animation (16 careers): Actor, Animator, Broadcast Technician, Camera Operator, Film Director, Film Producer, Film and Video Editor, Lighting Technician, Makeup Artist, Screenwriter, Storyboard Artist, Visual Effects Artist

• Games and digital content (11 careers): Esports Event Coordinator, Game Artist, Game Programmer, Game Tester, Level Designer, Social Media Specialist, Video Content Producer, Video Game Designer, Web Designer

• Music and audio (9 careers): Composer, Live Sound Engineer, Music Teacher, Musician, Sound Engineer, Talent Agent

• Writing and communication (15 careers): Author, Editor, Journalist, Proofreader, Public Relations Specialist, Sign Language Interpreter

• Technology, data, AI and cybersecurity (26 careers): Blockchain Engineer, Business Intelligence Analyst, Cloud Engineer, Computer Hardware Engineer, Computer Systems Analyst, Computer and Information Research Scientist, Data Scientist, Database Administrator, DevOps Engineer, IT Support Specialist, Information Security Engineer, Machine Learning Engineer, Network Administrator, Penetration Tester, Software Developer, Web Developer

• Robotics, automation and manufacturing (11 careers): Automation Engineer, CNC Programmer, Controls Engineer, Electro-Mechanical Technician, Industrial Machinery Mechanic, Manufacturing Engineer, Mechatronics Engineer, RPA Developer, Robotics Engineer, Robotics Technician

Sources

• O*NET Database, U.S. Department of Labor, Employment and Training Administration (USDOL/ETA), used under CC BY 4.0. Tasks, related titles and tools come from ONET 30; skills, education levels, interests and job zones come from earlier ONET releases. See DATA-SOURCES.md for the required attribution.
• U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics (OEWS): national pay percentiles (May 2025 where published, otherwise May 2024) and state medians (May 2024). Each pay section shows its date.
• National Center for Education Statistics: CIP 2020 to SOC 2018 crosswalk, linking college programs to occupations.
• Role-specific profiles: researched in October 2026 from BLS, professional associations and credential bodies. Salary figures link to their sources and were spot-checked against BLS pages. The Astronaut profile uses NASA's published requirements.
• Links to current information: every career page links to O*NET OnLine, BLS wage data, the BLS Occupational Outlook Handbook, CareerOneStop (certifications, state licenses, local training), Apprenticeship.gov and job searches (Indeed, LinkedIn, USAJOBS).

What has been tested

Tested in a headless Chromium browser on October 25, 2026:

• Quick match: three simulated students answered "yes" only to scenarios fitting their interest. Robotics: 11 questions → Controls Engineer, then Mechatronics Engineer, Drone Pilot, Robotics Engineer, Astronaut. Animals: 8 questions → Zoologist, then Animal Enrichment Specialist, Animal Trainer, Veterinary Technician. People and leadership: 11 questions → Executive Coach, then Organizational Development Consultant and Academic Advisor. Back, Skip, Change my picks, the step-by-step path for the top match and other matches, opening a full profile, and returning to results all worked.

• No JavaScript errors while loading or using any screen.

• The quiz: multi-select on every step, going back to earlier steps keeps selections, unselecting a parent clears its children, the live match count updates, and selections and the note persist after reloading.

• Results: "why it showed up" explanations, family filter, sorting, removing a selection from the results page.

• Career library search and family filter (all 19 families plus "All paths"), the role-specific filter, save, compare (up to three), and the About page.

• Career pages for both profile types, including the state pay picker and the Astronaut NASA source link.

• Phone width (390 px): no horizontal scrolling. All quiz options are real buttons with aria-pressed, so they work with a keyboard and screen readers announce their state.

• tools/validate.py checks that every O*NET code exists, every tag is a real quiz scenario, there are no duplicate titles, and every scenario leads to at least two careers.

Regression examples. These recreate the five checks from the original prototype. Counts differ because this version has a different catalog (319 cards instead of 228) and its own tags:

Path	Prototype	OpenPath	Why it differs
Technology → Robotics → "Troubleshooting a robot that stopped working"	14	10	Only careers whose day-to-day work includes fixing robots or automated machines are tagged (for example Robotics Technician, Controls Engineer, Industrial Machinery Mechanic, Astronaut). Broader engineering roles are reached through other robotics scenarios.
Creative → Film and TV → "Writing a screen story and seeing it filmed"	7	9	Includes storytelling roles beyond film crews, such as Narrative Designer (games) and Author.
Analysis → Finding patterns in data → "Spotting an unusual pattern in a dataset"	7	14	Many careers across families analyze data (for example Epidemiologist, Fraud Analyst, Radiologist, Climate Change Policy Analyst), so this scenario intentionally spans families.
Hands-on work → Using tools → "Repairing equipment that broke down"	10	13	The larger trades, aviation and robotics catalog adds more repair roles.
Technology robot-troubleshooting scenario + Teamwork "Building a product with a team of specialists"	26	24	Union of the two scenarios. Careers matching both (for example Robotics Engineer) rank first.

Not tested: real screen readers (only the structure was checked), every browser (only Chromium), performance on very old phones, printing on every printer, and outcomes with real students. The matching has not been validated as a career assessment.

Known limitations

• Coverage: 319 curated titles, not every job. Many specialty titles share one official occupation.
• Salary: BLS pay describes broad occupations. A specialty title's real pay may differ, and pages say so. Pay varies by location, employer, experience and time.
• Licensing and credentials: listed as things to look into, not legal requirements. Rules vary by state and change. Students should confirm with the credential body or their state licensing board (each page links to CareerOneStop's license finder).
• Estimates: a few newer O*NET occupations have no published survey data yet. Their skills, interests or education come from the most similar occupation, and the page says so.
• Shared guidance: family-level advice is general and may not fit every title in that family.
• Matching: hand-tagged scenarios plus simple word matching; not a validated assessment.
• Data dates: O*NET and BLS data were current as of October 2026; refresh them periodically (see below).

How to add careers or update data

All content lives in tools/catalog.py and data/.

Add a career

1. Find its official O*NET-SOC code at O*NET OnLine (for example 15-1252.00).
2. Add a line to CAREERS in tools/catalog.py:

   ("Game Audio Designer", "music", "27-4014.00", "Creates sound effects and music systems for video games.", "mus_sfx anim_level ind_create"),
   

   Fields: title, family id (one of the keys in FAMILIES), O*NET code (or "" if none), a one-sentence description, and space-separated scenario ids from INTERESTS/STYLES that this career fits.
3. Run python tools/validate.py and fix anything it reports.
4. Run python tools/build.py to rebuild index.html.

Add role-specific research for a career: add an entry with the same title to data/researched-profiles.json (copy an existing entry's structure: description, also called, and a profile with day, environment, pay with source, skills, certifications, learning, high school classes, majors, networking, ways in, path and branches). Rebuild. The page switches from shared guidance to role-specific details automatically.

Add or edit quiz questions: edit INTERESTS or STYLES in tools/catalog.py. Keep ids unique, tag careers with any new scenario ids, then validate and rebuild.

Update family guidance: edit FAMILIES in tools/catalog.py.

Refresh official data: data/official-occupations.json holds the processed ONET, BLS and NCES data (occupations, titles, tasks, skills, education, pay and majors). Regenerate it from new ONET and BLS releases, keep the same structure, then validate and rebuild.

Files

Path	What it is
index.html	The app (open this)
openpath-prototype.html	The earlier OpenPath prototype, kept for reference
Hackathon Submission.md	Hackathon submission write-up
tools/catalog.py	Quiz questions, career families and career cards
tools/validate.py	Catalog checks
tools/build.py	Builds index.html
tools/template.html	App design and code (Quick match, detailed quiz, results, library, career pages)
data/official-occupations.json	Processed O*NET, BLS and NCES data
data/researched-profiles.json	Role-specific research

---

© 2026 [Your Name]. All rights reserved. See LICENSE.md. Third-party data is credited in DATA-SOURCES.md.