HTCH | Devpost

Landing page
Network game subpage
Resources subpage
Mentors subpage
Network Event Finder subpage
Profile Subpage
Messages Interface
Login Page
Saved Mentors subpage

Inspiration

University students and newly graduated students entering the workforce often lack experience in maximizing their use of LinkedIn to connect with others in the same field, particularly with potential mentors, network, and discover tailored opportunities. Women are at the greatest disadvantage because they already face fewer role models in their fields, and it’s harder to find mentors who share their background or experiences. On top of that, biases and underrepresentation make networking feel intimidating, which means they’re often left out of the same opportunities their peers can access more easily.

What it does

It’s an AI-powered application designed to work hand-in-hand with LinkedIn, transforming professional networking into something structured, supportive, and even fun:

Mentorship Matching: HTCH suggests potential mentors within the user’s industry, surfacing connections that align with their career goals and background.

Tailored Opportunities: Beyond people, HTCH highlights relevant career-advancing opportunities, from internships to industry programs.

Networking Events: The platform helps users discover events where they can expand their professional circles in meaningful ways.

AI Assistant: Whether it’s drafting outreach messages, preparing for coffee chats, or following up, HTCH provides personalized coaching and communication support.

Game Mode: To make networking less intimidating, HTCH gamifies the experience. Users take on “missions,” earn points and badges, and climb leaderboards — turning professional growth into an engaging challenge.

In short: LinkedIn is the directory. HTCH is the guide, coach, and motivator that helps you turn those connections into lasting relationships and real career opportunities.

How we built it

Backend (Flask + MongoDB)

We used Flask as the main web framework to handle routing, authentication, and server-side rendering.

MongoDB was chosen as our database to store user accounts, mentor/mentee profiles, and progress from the game feature.

Session handling was implemented using Flask-Session, with secure password hashing from Werkzeug to protect logins.

Frontend (HTML, CSS, JavaScript)

The site uses standard HTML templates with Jinja2 (Flask’s template engine) to dynamically render pages such as login, register, dashboard, and game.

We built custom CSS and JavaScript for interactive features like the networking game, leaderboard, and UI responsiveness.

The app integrates with LinkedIn via profile data scraping / connection suggestions, presented through a clean, user-friendly interface.

AI Integration

We connected an AI assistant (via Snowflake Cortex / external API) to help users draft outreach messages, practice networking conversations, and receive career guidance.

This was encapsulated in a helper module (snowflake_client.py) that makes secure API calls to the model.

Game Feature

To gamify networking, we added a points system and leaderboard. Users earn points by completing networking missions, attending events, and connecting with mentors.

This logic is contained in the game module, with templates (game.html) and supporting scripts (gamescript.js).

Deployment & Collaboration

The project is version-controlled with Git and GitHub to collaborate seamlessly as a team.

For deployment, we prepared the app to run with Gunicorn (production WSGI server) so it could be deployed on cloud platforms like Heroku or Render.

Challenges we ran into

Merging our data. Integrating MongoDB and the Snowflake API. Connecting the LinkedIn sign in. Building a website. Having the game load updates.

Accomplishments that we're proud of

“Find a Mentor” directory with fast filters, save-via-flag, and Message {Name} CTA. Messages with left/right chat bubbles and the HTCH Chatbot for drafting intros. Resources & Events streamlined with clear Learn more / Apply actions and distinct event images.

What we learned

Locking a stable baseline before adding features prevents regressions. Good contrast/spacing/motion are essential for readable dark mode. Lightweight client state (filters, saves, threads) delivers smooth UX. How to integrate APIs and databases/

What's next for HTCH

Smarter filters and tags with tailored picks based on your interests and saves. Richer event/opportunity cards with key details, “save for later,” and gentle deadline nudges. Mobile refinements, an accessibility sweep, simple privacy controls, and light-touch moderation.