Inspiration
I built hyProX because most productivity tools feel made for average workflows, not for people trying to do a lot and move fast.
A lot of the time, the real questions are:
- Which opportunity is actually worth my time?
- Where do I stand compared to people I admire?
- What should I be doing next if I actually want to grow fast?
hyProX is my attempt at building something more strategic than just another “stay productive” app.
What it does
hyProX is a web app for ambitious, high-performing people who want to make better decisions about their growth.
- War Room: personal dashboard for profile, opportunities, benchmarks, progress, and next move
- Opportunity Intel: paste a hackathon / fellowship / conference / grant / event link and get key info, deadlines, fit, and whether it’s worth pursuing
- Signal Map: shows profile strength, missing proof, and how clear your overall story is
- Benchmark Circle: compare yourself with people you admire to identify meaningful patterns, gaps, and signals
- Wall of Works: execution layer for keeping top priorities visible and maintaining momentum
How I built it
I kept the stack modern but practical:
- Frontend: React + TypeScript, Vite, Tailwind CSS, Framer Motion, React Router, React Hook Form + Zod, Zustand
- Backend / Data: Firebase Authentication, Firestore, Firebase Storage, Cloud Functions
- AI: Google Gemini for most structured analysis and extraction, plus Qwen3-32B via Featherless.ai
Challenges I ran into
The two biggest challenges were:
- Making AI output usable: turning model output into something structured, relevant, trustworthy, and actually useful
- Controlling scope: the idea could have grown too big, so I kept bringing it back to the core loop: analyze → decide → act
Accomplishments that I’m proud of
I’m proud that hyProX feels like a real product and not just a collection of AI features.
Some things I’m especially happy with:
- bringing profile analysis, opportunity analysis, comparison, and execution into one place
- getting Firebase-based persistence and data flow working smoothly
- building something that feels genuinely useful, not just hackathon-useful
What I learned
A big takeaway for me was that AI is only valuable when it fits naturally into a real workflow.
What matters is not just that a model gives an answer, but whether it:
- helps you make a better decision
- saves you time
- gives you more clarity
I also got reminded that reliability matters just as much:
- auth
- state management
- redirects
- clear UI states
- understandable errors
Those details make a huge difference in whether a product actually feels solid.
What’s next for hyProX
Next, I want to build a much more personal accountability and strategy layer on top of this.
Something that doesn’t just analyze opportunities and profiles, but actually understands:
- what you’re trying to become
- what you’re avoiding
- where you tend to lose momentum
- what you should focus on next
Sponsor Credits
- Featherless.ai: Used as part of the project’s AI/model workflow
- gen.xyz: The project’s
.xyzdomain is registered through gen.xyz


Log in or sign up for Devpost to join the conversation.