Build FirstxGirls Who Code

    Build Your Career Decision Engine

    Girls Who Code Upskilling Series

    March 3, 2026  |  7:00 – 8:30 PM ET  |  Bethany Crystal  |  buildfirst.ai

    Quick Links

    Claude
    AI chat interface
    Claude Code
    Terminal-based AI coding
    Cursor
    AI code editor
    VS Code
    Code editor

    Today's Build

    1
    Welcome & Intros
    7:00 – 7:10 PM
    2
    Design Thinking with AI ("POP-IT")
    7:10 – 7:20 PM
    3
    AI Power Hour: Build Your Career Decision Engine
    7:20 – 8:10 PM · Sprint 1: Context Pack · Sprint 2: Decision Logic · Sprint 3: Frontend
    4
    Flow State Building: AI-Native Careers + Q&A
    8:10 – 8:30 PM

    POP-IT Framework

    POP-IT is a design framework that helps you identify "AI-Shaped Problems" in your work.

    P
    Problem — What is the problem you are trying to solve?
    O
    Output — What do you want to see from the AI when you're done?
    P
    Prompt — What instructions does the AI need to accomplish the task?
    I
    Input — What context or documentation can you share with the AI?
    T
    Test — How will you know it's working?

    Framework established in partnership with Decoded Futures

    Sample Context Pack: Nancy Drew

    Don't have your own resume handy? Use Nancy Drew's files to practice!

    ResumeBioCareer Goals
    SPRINT 1

    Show the AI Who You Are

    1

    Build Your Context Pack

    Create a folder on your computer called career-decision-engine. Add a sub-folder called context-pack and add 3-5 files about yourself:

    • Your resume
    • A cover letter or personal bio
    • Your transcript or portfolio
    • Your personal website content
    2

    Define Your Goals

    Create a new file called career-goals.md and fill in this template:

    My Career Goals

- Time Horizon (ie: 6-12 months):
- North Star (1 sentence):
- Top 3 Skills I Want to Build:
- Industries or Problem Areas I'm Curious About:
- Non-Negotiables (2):
- "In 6 months, I'd be happy to have _____"
- "If I had to choose, I'd prioritize _______ over ________"
    3

    Open Your Folder in Cursor or VS Code

    Open your career-decision-engine folder in your code editor, then open the terminal and type claude to start Claude Code.

    4

    Ask Claude Code to Summarize Your Context

    Paste this prompt to create your profile:

    Read all files inside the context-pack folder.

Synthesize them into a structured professional profile in JSON format with the following fields:
- strengths
- core_skills
- recurring_themes
- values
- target_directions
- constraints
- growth_edges
- risk_factors

Be concise but specific. If something is unclear or missing, add a field called questions_for_me.

Return only valid JSON.

Save the file as profile.md
    SPRINT 2

    Build Your Decision Engine

    1

    Define Your Evaluation Logic

    Tell Claude Code how to evaluate opportunities for you:

    We are building a Career Decision Engine. Start a file called decision-engine.md to store the logic for our evaluation.

Using profile.md as context, I want you to help me evaluate any job opportunity I paste in. When I provide a job description or a career opportunity, return:
- fit_score (0-100) compared to my background
- alignment_breakdown (skills, direction, values, constraints)
- tradeoffs
- missing_information_questions
- recommendation (yes / no / investigate)
- 7_day_action_plan
- what_to_ignore

Rules:
- Be honest about uncertainty.
- If key information is missing, recommend "investigate."
- Keep reasoning concise but specific.

Ask me any follow-up questions before we build.
    2

    Answer Claude's Questions

    Claude Code will ask you clarifying questions about your preferences, dealbreakers, and priorities. Answer them thoughtfully — this calibrates your engine!

    3

    Test It on a Real Job

    Find a job posting you're curious about and paste it into Claude Code. Review the evaluation — does it match your intuition?

    Example test prompt:

    Test it on this job: [paste job URL or description here]

    Security Note

    • Paid Claude accounts don't use your data for model training
    • If using free tools: remove phone number, address, and other PII from your resume
    • Never expose API keys in frontend code
    • Local builds are not the same as deployed apps
    • Today = local experiment, not production software
    SPRINT 3

    Build a Frontend (Make It Repeatable)

    1

    Tell Claude Code to Build Your App

    Use this prompt to generate a simple web interface:

    Build a local web app that evaluates a job description using my profile.md and decision-engine.md.

Ask me questions about my goals before we build.

Call out any security questions along the way.

I want this to be something simple that I can build in under 1 hour.
    2

    Answer Architecture Questions

    Claude Code will ask about tech stack, input methods, and priorities. Good defaults for a quick build:

    • Node + Express (simple backend)
    • Call Claude API for evaluation
    • Paste text input (simplest)
    3

    Review and Approve the Plan

    Claude Code will show you the architecture before building. Review it, then approve to start coding.

    4

    Run Your App Locally

    Follow Claude Code's instructions to start your app (usually npm run dev or node server.js), then test it in your browser!

    API Key Management

    • API keys live in .env and should never be committed or shared
    • If hosting locally, your API keys can stay in .env
    • If deploying publicly, save API keys as environment variables in your hosting platform
    • If API keys are compromised, rotate them immediately
    • Set spending limits on your API account

    Tips for AI Coding

    • Make sure you know the problem and the audience before you build
    • Use Gemini, Claude, or ChatGPT to iterate on your prompt first, then make the app
    • Run your app locally before deploying publicly
    • Get the AI to perform reliably before introducing websites or app interfaces
    • If you hit a bug or don't know what to do, ask the AI for help!

    What Career Opportunity Are You Evaluating?

    Resources

    Build First
    AI learning lab
    Claude Docs
    Official documentation
    Girls Who Code
    Community