Trailhead — AI-matched opportunities. Your skill, your path.

Inspiration

A few weeks ago, my college was selected to host a TCS AI Hackathon. I was excited to participate — but I never got the chance. Selection wasn't open; it was decided by a teacher, who only picked a handful of "high priority" students. I wasn't one of them.

That moment stuck with me. It wasn't about not being good enough — I never even got a shot to prove it. And I knew I couldn't be the only student this happens to. Every year, thousands of students miss out on hackathons, scholarships, and internships simply because they didn't know the opportunity existed, or because access was gatekept by someone else's decision.

So instead of waiting to be picked again, I decided to build something that lets students find and go after these opportunities themselves — no gatekeeper required.

What it does

Trailhead is an AI-powered web app that matches students to real hackathons, scholarships, and internships based on who they actually are — their skills, interests, experience level, and preferences.

Students choose one of three paths:

Hackathons — based on interests, experience level, and whether they want to go solo or find a team Scholarships — based on education level, field of study, and what matters most to them (merit, need, field-specific, etc.) Internships & Jobs — based on their skills, desired role, and preferred company type

After answering a few quick questions, the AI analyzes their profile against a curated database of real opportunities and returns the top 5 matches — each with a match confidence score and a short, specific explanation of why it fits them. Every result links directly to the real registration or application page.

How we built it

Trailhead has two layers working together: a frontend and an AI-powered matching engine.

Frontend: The entire interface — the category forms, chip-based selectors, scanning animation, ranked result cards, and saved-items list — is built with plain HTML, CSS, and JavaScript, no frameworks. I designed it around a "trailhead / signpost" theme: the idea that every student's path starts somewhere, and this app is meant to be that starting point.

AI matching engine: The actual intelligence behind Trailhead comes from Anthropic's Claude AI model. Here's the flow:

The student fills out a short form (interests, skill level, preferences — different fields depending on the category they pick). Their answers are converted into a plain-language profile. That profile, along with a curated list of real opportunities for that category, is sent to the AI. The AI analyzes the fit between the student's profile and every opportunity in the list, then returns the top 5 matches — ranked, scored, and with a short, specific reason for each one. The results are rendered live as ranked cards, each linking straight to the real registration or application page.

So nothing about the ranking or reasoning is hardcoded — only the underlying list of opportunities is a curated dataset (to keep the project realistic to build solo in the time available). Every match, score, and explanation is generated live by the AI based on that specific student's answers.

This is my first time building a full AI-powered web app. My background is in C++, SQL, and Node.js, and this project pushed me to learn how to actually integrate an AI model into a working product.

Challenges we ran into

. Getting the AI to return consistent, structured results so they could be displayed as clean, ranked cards . Designing three different forms that still felt like one cohesive product . Scoping the project so it was realistic to finish solo in the time available

Accomplishments that we're proud of

. Going from "I didn't get picked for a hackathon" to shipping a working AI product in the same season . Building something that could genuinely help other students in my position . Learning AI integration from scratch, solo, without a team

What we learned

How to actually connect an AI model into a real product — not just chatting with it, but making it the core decision-making layer of an app How to design a prompt that gets consistent, structured output from AI (instead of random freeform text) so it could be displayed as clean, reliable results every time That scoping matters more than ambition — cutting the live-scraping idea and using a curated dataset instead is what actually made this finishable solo in the time available That a genuine personal problem makes for a stronger project than a "cool idea" picked at random

What's next for Trailhead — AI-matched opportunities. Your skill, your path.

.Expanding the curated database into a live, continuously updated feed of opportunities .Adding deadline reminders for saved opportunities .Adding a resume/profile upload for even more personalized matching .Opening it up so other students facing the same problem can actually use it

Built With

• ai
• anthropic
• claude
• css
• html
• javascript