This is a submission for the GitHub Finish-Up-A-Thon Challenge
What I Built
FoundrGeeks — a global platform that eliminates founding team
risk for startup founders worldwide.
The single biggest risk you face as a startup founder isn't your
product or your market. It's your founding team. More startups
die from co-founder conflict and misalignment than running out of
money. FoundrGeeks solves this.
It connects tech visionaries, developers, designers, product
managers, and founders — from across the world, empowering them
to forge meaningful collaborations, discover compatible co-founders,
and build the next generation of transformative startups.
Core Features:
- 🤖 AI-Powered Talent Matching — Uses NVIDIA Text Embeddings and LangChain to perform cosine similarity vector searches that mathematically calculate co-founder compatibility based on bios, skills, and goals
- 🚀 Project Showcase Board — Founders can pitch their startups and recruit the exact roles they need
- 💬 Real-Time Discussion Forums — A chatty, Facebook-style community space for collaboration, feedback, and networking
- 🔔 Smart Notifications — Real-time alerts for matches, replies, and new connections
Tech Stack: Next.js, Node.js, Express, MongoDB, NVIDIA Text
Embeddings, LangChain, JWT Auth
🔗 GitHub: https://github.com/Isreal-Oparanti/Launchpad
Demo
The Comeback Story
This project had been sitting unfinished in my GitHub repository
for months. The foundation was there, user authentication, basic
profiles, a project board, and basic project matching algorithm I built, it was getting to large to push across the finish line alone.
The AI matching engine there but needs improvement.
The discussion forums new some UI/UX finish. The UI was inconsistent and
incomplete. The database had no real data to demo with.
Here is what I added to finally finish it:
- ✅ Built the entire AI matching engine from scratch using NVIDIA Text Embeddings and cosine similarity vector search
- ✅ Designed and built the full Discussion Forums feature with threaded replies, real-time comment counting, and delete functionality
- ✅ Completely overhauled the UI, profile pages, talent cards, project cards with real logos and cover images
- ✅ Built a robust database seeding system with realistic demo users and projects so the platform feels alive
- ✅ Fixed multiple UI bugs including avatar rendering, logo loading, and mobile navigation glitches
- ✅ Improved the real-time notification system
A project that felt impossible to finish alone went from 40% done
to 98% done with the help of github copilot.
My Experience with GitHub Copilot
GitHub Copilot genuinely changed how I work. It wasn't just
autocomplete, it was a true pair programmer that kept me moving
when I would have otherwise been stuck.
Specific moments where Copilot made the difference:
The AI Matching Engine — Implementing cosine similarity vector
search with NVIDIA Text Embeddings was the most technically
challenging part of this project. Copilot helped me understand the
architecture, suggested the right LangChain patterns, and helped
me debug edge cases that would have cost me days alone.
Discussion Forums Backend — Wiring up the full CRUD logic for
threaded replies, including real-time count updates, pagination,
and the delete permission system, was completed in a fraction of
the time it would have taken me. Copilot suggested the mongoose
query patterns and notification logic almost instantly.
Database Seeding Scripts — Copilot helped me write realistic,
structured seed scripts that populate the platform with believable
demo users and projects — something I kept skipping because it felt
tedious. Having Copilot generate the boilerplate made it fast and
even enjoyable.
Debugging — Several times I hit UI glitches (logo loading
failures, avatar rendering issues) and Copilot helped me trace
the root cause quickly, suggesting the exact fix I needed.
The honest summary: tasks that would have taken me days took hours.
GitHub Copilot gave me the momentum and confidence to stop staring
at an unfinished repository and actually ship.
Top comments (5)
Matching with AI?
Brilliant 👏
Yes primary vector based matching important detail like location, skills compliment, track records, projects needs etc
There's more room and high chance you'll find what you're looking for
Awesome 👌
How is this better that YC matching tool, solvearn etc?
Don't get it mixed up.
This isn't just a tool.
FoundrGeeks is a system built specifically for:
networking
- meaning you discover people around you with shared interests.
building your team
- Finding cofounder, skilled professionals and investors
Hiring
- You get go hire efficient individuals from the same network, whether as an early stage startup or a skilled professional looking for a dream to be a part of.
YC marching tool dont have any of those aforementioned.