Career Compass

🚀 Overview Career Compass helps global talent find legitimate, high-signal job opportunities and avoid wasting time on low-quality or fake listings.

Instead of endlessly applying into the void, users get:

ranked opportunities legitimacy scoring interview likelihood prediction AI-generated career insights a voice-enabled career coach (emotional + tactical) warm referral intros to people already inside target companies Built with Google Gemini 2.5 Pro (via Vertex AI) for the Best Use of Gemini hackathon category — where Gemini is the decision engine at every step, not a cosmetic add-on.

⚠️ Problem Modern job seekers face three core issues:

🔴 Job boards are noisy and unfiltered 🔴 Remote jobs are often fake, outdated, or misrepresented 🔴 AI hiring systems auto-reject candidates before human review As a result, skilled candidates waste weeks or months applying blindly with no signal. Diaspora and Global South talent are hit hardest — less network visibility, and more exposure to region-specific scams.

💡 Solution Career Compass introduces an AI career intelligence layer that helps users decide:

“What should I apply to first — and what should I ignore?”

We turn job search from volume-based to signal-based decision making. Every job has to answer three questions before it earns your time:

Is this real? · Is this worth my time? · Do I have a chance?

🧠 Core Features (MVP)

1. Resume Intelligence Upload resume or paste profile Gemini extracts (in under 3 seconds, no forms): skills seniority level role intent strengths & gaps origin city + years of experience
2. Job Scoring Engine (Gemini-Powered) Each job is evaluated across:

🎯 Match Score (fit to resume) 🛡️ Legitimacy Score (trust / scam / credibility filter) 📈 Interview Likelihood Score 🌍 Remote Authenticity Score ⚡ Apply Priority Rank

1. “Why You Might Get Rejected” (Key Innovation) Gemini explains:

missing skills mismatch signals ATS / screening risks role misalignment This turns the system into a career advisor, not just a job board.

1. Ranked Opportunity Feed Users see:

Top 3–5 best opportunities only Clear scoring breakdown Apply / Skip recommendations Priority ordering based on success probability Badge Trigger ⚡ Apply Now High match · Legitimacy ≥ 85 · interview High ◷ Maybe Later Solid match but high competition or medium signal ⊘ Skip Legitimacy < 50 or strong scam / anomaly flags

1. Nia — AI Career Coach A voice-enabled coach powered by Gemini with full access to the user's live profile — not a generic FAQ bot, it knows who you are.

Pairs tactical strategy (what to apply to, how to position) with mental-health-aware pacing to fight burnout Voice-to-text input for the hard days when typing is too much Turns a natural conversation into live search filters (cognitive style, workstyle, location, salary goal, target role)

1. Warm Referral Network Surfaces community members already referring at companies you fit Gemini drafts a personalized intro DM for each one Warm intros convert roughly 3× a cold apply 🧪 How It Works Resume Upload ↓ Gemini Profile Extraction ↓ Job Dataset Ingestion ↓ Gemini Evaluation Layer ↓ Scoring Engine:
   • Match Score
   • Legitimacy Score
   • Interview Likelihood ↓ Ranking System ↓ Top Opportunity Feed 🏗️ Architecture (Built Entirely on Google Cloud)

Layer Google service Job Intelligence Gemini 2.5 Pro / Vertex AI Every scoring, extraction, writing & coaching decision Matching Vertex AI Vector Search Embeds profiles + jobs, finds similarity at scale Auth Firebase Auth One-tap Google sign-in, no passwords Database Firestore Real-time store: profiles, jobs, applications, referrals Files Firebase Storage Resumes + photos, compressed for slower mobile networks Jobs pipeline Cloud Functions + Cloud Scheduler Scrape employer APIs every 6h, vet via Gemini Hosting Cloud Run + Firebase Hosting Auto-scaling app, global edge caching CI/CD Cloud Build Push to GitHub → live in under 3 minutes 🧰 Tech Stack Google Gemini 2.5 Pro via Vertex AI (core intelligence engine) Vertex AI Vector Search (matching) Firebase Auth · Firestore · Firebase Storage Cloud Functions + Cloud Scheduler (job pipeline) Next.js on Cloud Run · Firebase Hosting (frontend) Cloud Build (CI/CD) Python / FastAPI (scoring services) JSON / sample job dataset (hackathon mode) 🔥 What Makes This Different Most job tools:

optimize resumes auto-apply to jobs or match keywords Career Compass:

Optimizes decision-making under uncertainty

We don’t just tell users what jobs exist. We tell them:

which jobs are real which jobs are worth their time which jobs they are most likely to actually get 🧭 Key Innovation: Legitimacy Score A Gemini-powered trust filter that evaluates:

company credibility signals job clarity and structure hiring realism scam indicators (including African-market red flags: missing company site, unrealistic salary, suspicious apply URL) outdated or duplicate postings High-confidence legitimate listings go live automatically; only edge cases reach a human reviewer. This helps users avoid low-quality or fake opportunities.

🌍 Vision We believe the future of work is:

distributed AI-filtered referral-driven trust-based Career Compass builds the signal layer for global opportunity access.

"By the time a job seeker clicks apply, Gemini has already read their resume, scored the job, written their cover letter, found their referrer, and removed every obstacle between them and an opportunity."

🏁 Hackathon Goal Demonstrate how Gemini can be used not just for chat or content generation, but for:

🧠 real-world decision intelligence in workforce mobility

📽️ Demo Flow Upload resume AI extracts profile Jobs are evaluated using Gemini System returns ranked opportunities User sees: match score legitimacy score interview likelihood “why you may get rejected” Nia (voice coach) gets the user through the hard days and drafts a warm referral intro 🛠️ Status MVP built for hackathon demo

Future expansion:

referral graph network real job APIs diaspora signal matching system recruiter integration layer Built with ❤️ for global talent · by Afro Adventurer · Build54 Nairobi 2026

Built With

• typescript