GitHub Finish-Up-A-Thon Submission: Reviving Pathfinder AI
This is a submission for the GitHub Finish-Up-A-Thon Challenge
What I Built
Pathfinder AI is an AI‑powered career guidance tool that helps students and job‑seekers discover personalized career paths.
The "Before"
The project was initially a basic prototype with:
- Only English language support
- No video introduction
- Only ran locally, not deployed online
- Basic UI without proper guidance
The "After"
The revived version now includes:
- Full multilingual support (English, French, Spanish)
- Narrated video introduction using AI voiceover
- Live deployment on Streamlit Cloud
- Global Security Shield badge and transparent pricing in the sidebar
- Comprehensive README and documentation
How I Used AI and GitHub Copilot
I used Groq's Llama 3.1 model to power the AI career coach, and my own AI voice translator to generate the narrated video introduction. GitHub Copilot helped me quickly refactor the multilingual UI components.
Links
GitHub Repository: https://github.com/Deslandes1/revive-pathfinder-ai
Live Demo: https://ibm-s-global-ai-builders-challengecontest-2026-4plivmvuiyykuy4.streamlit.app/
Demo Video: https://www.dropbox.com/scl/fi/w5c4hxj0gldttd7ks3vcq/Pathfinder.mp4?rlkey=bqx234mre5bj30fwydruu4pm6&st=wvjsmggo&dl=1
Top comments (4)
Can you explain to me how the E2E works in combination with Ai?
I’d also love to know more about your global security shield
Hi Johannes, great questions – thank you for asking.
Let me break it down.
The Global Security Shield is a lightweight middleware that protects Python web applications (Streamlit, Flask, FastAPI, etc.) from common web attacks:
It works by scanning every user input (text fields, URL parameters, headers) against a set of regex patterns – both built‑in and custom. If a match is found, the request is blocked immediately, and a detailed log is sent to a central Security Dashboard. The dashboard shows real‑time alerts, registered apps, and threat logs.
Key features:
I built this shield to protect my clients’ applications – it is not a product for sale. It runs on every app I deliver, including the AI Customer Service Suite, Hospital Management System, and System Health AI Monitor.
In all my AI‑powered software, the shield sits in front of the AI. Here’s how:
So the shield does not rely on AI to detect attacks (that would be slow and unpredictable). Instead, it provides a deterministic, low‑latency first line of defence. AI then operates on trusted, clean data.
You are right to ask. In my dashboard, I display a “Secure channel active” badge. That badge refers to the TLS/HTTPS connection between your browser and the app – it is not end‑to‑end encryption of data stored on the server.
The shield does not currently implement E2E encryption of logs or user data. If you need true E2E (where only the client can decrypt data), that would require client‑side encryption and is not part of this shield. I plan to add optional E2E for sensitive logs in a future version – but today, it protects in transit and at the application level against injection attacks.
I hope that clarifies everything. The shield is a tool I use to keep my clients safe – not something I sell separately. Feel free to ask more.
Best regards,
Gesner Deslandes
Engineer‑in‑Chief, GlobalInternet.py
Thank you for your reply!
I did not expect such a long and insightful response.
I thought your project was very interesting :)))
Please know that you’re valuable and that you matter.
Thank you. I love coding and I build software on a daily basis.